AIAA JOURNAL

Vol 38 No 9 September 2000

Review of Propulsion Applications of Detonation Waves

K Kailasanathcurren

US Naval Research Laboratory Washington DC 20375

Applications of detonations to propulsion are reviewed First the advantages of the detonation cycle over theconstant pressure combustion cycle typical of conventional propulsion engines are discussed Then the earlystudies of standing normal detonations intermittent (or pulsed) detonations rotating detonations and obliqueshock-induced detonations are reviewed This is followed by a brief discussion of detonation thrusters laser-supported detonations and oblique detonation wave engines Finally a more detailed review of research duringthe past decade on ram accelerators and pulsed detonation engines is presented The impact of the early work onthese recent developments and some of the outstanding issues are also discussed

Introduction

T HE power of detonations has been well recognized For exam-ple it has been estimatedthat a 20-m2 detonation wave operates

at a power level equal to that received by the Earth from the sun1

The dif culty in harnessing this power ef ciently has been a majorstumbling block in the development of propulsion systems based ondetonations Although there are no practical propulsion systems (tothe authorrsquos knowledge) using detonations it has not been due tolack of effort As will be discussed therehave been serious attemptsat least since the 1940s

In principle detonations arean extremelyef cientmeans of burn-ing a fuelndashair mixture and releasing its chemical energy contentHowever detonations have been explored for propulsion applica-tions only for the past 50 years or so23 because of the dif cultiesinvolved in rapidly mixing the fuel and air at high speeds and initi-ating and sustaining a detonation in a controlled manner in fuelndashairmixtures Recently there has been a renewed interest in the appli-cation of detonations to propulsion and hence it is timely to reviewthe past work There have been several review papers (eg Refs 4ndash

8) in the past dealing with particular applications of detonations topropulsion Here an attempt is made to include the work reportedin those papers put them in the context of other related researchinclude additional studies and extend the review to more recentpapers

In this paper the status of propulsion applications of detonationsis reviewed First a cycle analysis is performed to show that the ef- ciency of a detonation cycle is close to that of the constant-volumeHumphrey cycle which is much more ef cient than the constant-pressure Brayton cycle characteristicof most conventional propul-sion systems Other advantages of detonations are also discussedThen a review of the early attempts to use detonations for propul-sion is presented After a brief discussion of the possible reasons forthe successes and failures of the early attempts more recent workduring the 1980s on oblique detonation wave engines is reviewedThis is followed by a detailed analysis of the ram accelerator in thedetonative mode and the pulsed detonation engine a topic of greatcurrent interest Finally some observations from the lessons learntin the past and their potential implications for further developmentof detonations for propulsion applications are presented

Why DetonationsAs mentioned before very rapid material and energy conversion

is a key featureof detonations This rapidburning or materialconver-

Presented as Paper 99-1067 at the 37th Aerospace Sciences MeetingReno NV 11ndash14 January 1999 received 4 June 1999 revision received21 January 2000 accepted for publication 22 February 2000 This materialis declared a work of the US Government and is not subject to copyrightprotection in the United States

curren Head Center for Reactive Flow and Dynamical Systems Laboratory forComputational Physics and Fluid Dynamics Code 6410 Associate FellowAIAA

sion rate typically tens of thousands of times faster than in a amecan lead to several advantages for propulsion such as more compactand ef cient systems Because of the rapidity of the process there isnot enough time for pressure equilibration and the overall processis thermodynamically closer to a constant volume process than theconstant pressure process typical of conventional propulsion sys-tems To illustrate this point three idealized thermodynamic cyclesare compared in Fig 1

For purposes of comparison the only process that is differentin the three cycles is the mode of energy conversion or heat addi-tion For the three cases heat is added at constant pressure con-stant volume or in a detonation Hence the three cycles have beenreferred to as constant pressure constant volume and detonationcycle respectively The amount of heat added is kept the same forthe three cycles In all cases the fuelndashair mixture is initially com-pressed adiabatically from 1 to 3 atm before heat addition Afterheat addition the products of combustion are expanded adiabati-cally to 1 atm Finally the system is returned to its initial state Thework done during the three cycles is obtained from the area enclosed(Fig 1) Because all processes except for heat addition have beenmaintained the work done or relative thermodynamic ef ciency ofthe three combustion processes can be obtained by comparing thethree areas For the ef ciency the work output is divided by theheat input which was set to be the same for the three cycles Thethermodynamic ef ciencies for the three cycles are 27 for con-stant pressure 47 for constant volume and 49 for detonationFrom Fig 1 and the given values we see that the thermodynamicef ciency of the detonation cycle is close to that of the constant vol-ume cycle The process itself is different with a decrease in speci cvolume and a signi cantly higher pressure being attained duringdetonations

One of the factors that could change the relative ef ciencies isthe amount of initial compression To illustrate this several suchcycles were computed with all factors held the same except for theamount of initial compression There is some change in the relativevalues but in all of these cases the ef ciency of the detonation pro-cess is close to that of the constant volume process and signi cantlybetter than that of the constant pressure process Other factors thatcould affect the relative ef ciencies are the amount of energy con-version (heat addition) and the rates of initial compression (if any)and nal expansion These do not drastically alter the results justpresented

It is important toemphasizethat theprecedingcomparisons are foridealized thermodynamic cycles for systems operating in a steadystate and are not representative of any particular propulsion sys-tem As will be discussed attempts at developing engines usinga steady or stabilized detonation wave have been less successfulthan those taking advantage of the unsteady aspects of detonationwaves

Several other advantages have been stated for using detonationsin propulsion devices and these will be brought up as appropriatewhen the different applications of detonations are discussed

1698

KAILASANATH 1699

Early ResearchAlthough basic studies and applications of detonation have been

undertaken for a very long time speci c references23 to propulsionappear in the literatureonly in the 1940s Even at this early time bothstanding (or stabilized) and unsteady (intermittent)detonations wereexplored In the work of Hoffmann2 both gaseous (acetylene) andliquid (benzene) hydrocarbon fuels were employed with oxygenIntermittent detonation appears to have been achieved but attemptsto determine an optimum cycle frequency were less successful Thedevelopment of the concept of pulsed detonation engines (PDEs)has been traced back to this pioneering work in a number of papersThe proposals of Roy3 inspired further work in France on the de-sign of systems to stabilize combustion in supersonic ows (egRef 9) Soon work was also begun in the United States Bitondoand Bollay10 conducted an analytical study that indicated that a

Fig 1 Comparison of idealized thermodynamic cycles for constantpressure constant volume and detonation modes of combustion

Fig 2 Schematic of a standing detonation wave at the exit of a nozzle(from Ref 14)

Fig 3 Schematic of an experimental setup for multicycle detonation studies (from Ref 16)

pulsating detonation engine can be helpful for helicopter propul-sion Gross11 Gross and Chinitz12 Nicholls et al13 and Nichollsand Dabora14 studied means to stabilize detonation waves for ap-plications to hypersonic ramjet propulsion The possibility of usingoblique detonations was also introduced15 More recent extension ofthese ideas will be discussed in the sections on ldquoOblique DetonationWave Enginesrdquo (ODWEs) and ldquoRam Acceleratorsrdquo

Stabilized Normal DetonationsFirst we brie y look at some of the early experiments to stabi-

lize detonation waves A schematic of an experiment is shown inFig 2 (from Ref 14) Cold hydrogen gas injected at the throat ofa convergentndashdivergent nozzle mixes with the high-pressure high-temperature air owing through the nozzle Because of the shortresidence time and rapid drop in temperature combustion does notoccur within the nozzle The nozzle is operated highly underex-panded so that further expansion occurs outside resulting in a com-plex system of shock waves a key feature of which is the Machdisk a nearly normal shock The conditions behind this Mach diskare such that ignition and energy release occur just behind it If theenergy release is closely coupled to the shock wave a detonationwill be established Whether or not a detonation is established willdepend on the induction time behind the Mach disk which in turnwill depend on the pressure temperature and mixture compositionApplication of such steady-state detonations to propulsion were ex-plored and performance comparable to conventional ramjets werereported for appreciably higher ight Mach numbers15

Intermittent DetonationsConcurrent with their work on stabilized detonation waves

Nicholls et al16 also explored the concept of intermittent (or pulsed)detonation waves for propulsion applications Both single-cycle andmulticycle operations with hydrogen and acetylene as fuels andoxygen and air as oxidizers were demonstrated The basic setupwas a simple detonation tube open at one end with coannular fueland oxidizer injection at the closed end as shown in Fig 3 (fromRef 16) Thrust fuel ow air ow and temperature measurementswere made over a range of operating conditions When the sparkplug was located 2 in (508 cm) or 5 in (127 cm) from the end ofthe mixing plane spasmodic ring was observed suggesting prob-lems with fuelndashair mixing However with the spark plug located 10in (254 cm) downstream periodic detonations for a range of mix-tures were reported For a hydrogenndashair mixture a speci c impulseof 2100 s was attained along with a cycle frequency of 35 Hz Theyalso presented a very simpli ed theoretical analysis that gave over-all results very much in agreement with their experimental data onhydrogen but less with the case of acetylene16 However they real-ized that the agreement was partly fortuitous because the measuredthrust-time history was signi cantly different from the theoreticalresult (Fig 4 from Ref 16) Note that they also attempted initiationfrom the open end but were not successful

A setup similar to that of Nicholls et al16 was constructed byKrzycki17 who used automotive spark plugs for ignition He demon-strated operation at 60 Hz with propanendashair mixtures but there issome doubt whether the device was operating in the detonative mode

1700 KAILASANATH

Fig 4 Comparison of theoretical and experimental results on timehistories of detonation tube acceleration for a 50 acetylenendashoxygendetonation (from Ref 16)

Fig 5 Possible con guration for a rotating detonation wave engine(from Ref 19)

or merely as a pulse jet engine due to the low initiation energies em-ployed His overall conclusion was that although thrust was possiblefrom such a device practical applications were not promising

Rotating DetonationAnother interesting concept that was explored was that of a rotat-

ing detonation wave rocket motor18 iexcl 20 Here gaseous fuel (hydrogenor methane) and oxygen were continuously injected into an annularcombustion chamber and ignited using spark plugs A unidirec-tional detonation wave was created after transition from a de agra-tion wave and the exhaust gases were expelled through an annularnozzle Such a system is shown in Fig 5 (from Ref 19) The accom-panying analysis indicated that the ideal performance is essentiallythe same as that of a conventional rocket engine as long as the aver-age chamber conditions and fuels used were the same Furthermoreif there was signi cant mixing between the burned and unburnedgases the performance could actually be degraded19 Further workwas done to extend the analysis to two-phase detonations20 and toprovide an explanation of some cases of shock-related combustioninstabilitiesobserved in liquid rocket motors21 However multicycleexperimental operation does not appear to have been achieved

Oblique Shock-Induced DetonationsThe work of Gross11 Gross and Chinitz12 Nicholls et al13

Nicholls and Dabora14 and Dunlap et al15 led to an extensive ex-

Fig 6 Schematic of a normal shock-induced combustion (from Ref 6)

Fig 7 Schematic of an oblique shock-induced combustion (fromRef 6)

perimental study using a Mach 3 tunnel at the Arnold Engineer-ing Development Center A detailed review of this work duringthe period 1959ndash1968 is available6 and hence only a brief discus-sion is given here The shock bottle experiments of Nicholls andDabora14 (Fig 2) were reproduced and a different con gurationwas developed In this con guration [Fig 6 (from Ref 6)] combus-tion takes place behind a normal shock generated by oblique shocksinduced by wedges It has been argued that a detonation does nottake place in this con guration because the normal shock wave isindependently generated by the wedges and is not directly affectedby the combustion6 The term ldquoshock-induced combustionrdquo was in-troduced to describe such phenomena Whether a detonation occursor not their work led to the realization that a normal shock-induceddetonation a ChapmanndashJouguet (CndashJ) detonation was not essentialfor propulsion6 They investigatedother con gurations [Fig 7 (fromRef 6)] and concluded that oblique-shock-induced combustion isexperimentally feasible This was a signi cant conclusion becauseit has led to the future development of the ODWE and to someextent to the ram acceleratorBoth of these propulsion applicationswill be discussed later in some detail

Detonation ThrustersIn the 1970s a substantial effort22 iexcl 25 was undertaken at the Jet

Propulsion Laboratory (JPL) to investigate the feasibility of usingdetonative propulsion for thrusters in the dense or high-pressure at-mospheres of solar system planets The primary reason for studyingthis concept was the potential for high performance for situationswhere the external pressures are high Under such conditions con-ventional thrusters become inef cient due to the lowering of thecombustor to ambient pressure ratio The JPL concept involved ll-ing nozzles with hydrogen helium nitrogen or carbon dioxide tosimulate planetary atmospheres and then intermittentlydetonating asmall quantity of condensed-phase explosives at the apex of the noz-zle Thrust is produced primarily from the momentum of the llergases set into motion by the blast wave and by the expansion ofthe detonation products Both theoretical analysis and experimentsshowed that a small quantity of explosives detonated within the noz-zle lled with gases was suf cient to produce high speci c impulseDifferent types of nozzles and gases of different molecular weightswere utilized With a short-plug nozzle there was a slight reductionin speci c impulse with increasing ambient pressure but the resultswere virtually independent of the molecular weight of the ambi-ent gases However with a long-cone nozzle there was a progres-sive increase in the speci c impulse with increasing ambient pres-sure for high molecular weight gases (carbon dioxide and nitrogen)and a decrease with increasing pressure for lower molecular weightgases (helium) A primary conclusion from the work at JPL was that

KAILASANATH 1701

detonation propulsion technology can be considered for producingboth large velocity changes as a main propulsion system or for gen-erating small velocity changes such as in attitudecontrol propulsionAlthough this can be considered to be one of the more successfulapplications of detonation to propulsion one must remember thatin this application the details of the detonation process itself ap-pear to be secondary and the working uid is a nondetonable gasmixture

Laser-Supported Detonations (LSDs)Another intriguing concept proposed in the 1970s was the use of

lasers to propel objects to low Earth orbit26 The basic idea behindthis propulsion concept is to use lasers to generate detonations anda high kinetic energy propulsive jet moving away from the objectbeing propelled The idea was not pursued very far due to the limitedlaser powers available then and the obvious atmospheric transmis-sion problems However with the planned development of largerlasers under the Strategic Defense Initiative program the conceptwas explored in some detail (for example see Refs 27ndash29) in thelate 1980s In principle a high controllable exhaust velocity can beachieved by a two-step process In the rst step the desired amountof propellant is formed as a gas layer near the surface and in thesecond step the laser energy is directly deposited in this layer The rst layer can be generated either by transpirationof a liquid througha porous surface or by a laser prepulse in a two-pulse system27 Ina two-pulse system a very low-intensity long-duration (on the or-der of seconds) laser prepulse is used to ablate enough gas fromthe surface of the solidliquid propellant to form a buffer gas layerA second high-intensity short-duration laser pulse irradiates thegas layer depositing its energy in the gas layer just before reachingthe propellant surface A shock wave then propagates out throughthe gas layer back toward the laser heating and accelerating thegas layer What is effectively a laser-supported detonation (LSD)is formed by the shock wave followed by the recombination zonebehind it28 In theory a high speci c impulse ( laquo 1000 s) can begenerated with simple propellants such as ice However the realproblems of building a large enough laser and safely transmittingenough energy to the desired location remain and to the best ofthe authorrsquos knowledge LSDs have not been conclusively demon-strated

Another concept using lasers to initiate detonations was proposedby Carrier et al30 Fendell et al31 and Carrier et al32 They con-ducted theoretical and experimental feasibility studies of a super-sonic combustor based on a stabilized conically con gured obliquedetonation wave (ODW) The stabilized detonation wave is formedby the interaction of a train of spherical detonation waves each di-rectly initiated by a brief localized deposition of energy from a veryrapidly repeated pulsed laser The laser is focused on a xed site inthe combustor where there is a steady uniform supersonic streamof detonable gases Initiation of an individual spherical detonationwave by a single laser pulse was achieved32 but further results havenot been reported

ODWEsWith the renewed emphasis on hypersonic ight vehicles in the

1980s extensive investigations were conducted of supersonic com-bustion concepts However these studies will not be discussed hereunless they speci callyinvoke detonations Various versionsof theseconcepts have been called by special names such as detonation-driven ramjets (dramjets)5 but these essentially involve stabilizedODWs and can be traced to the original concept proposed in the1950s by Dunlap et al15 Basically if a premixed fuelndashair mixture ows at a velocity greater than the CndashJ velocity then a normal det-onation wave cannot be stabilized The hope was that under theseconditions an ODW can be stabilized on a wedge-shaped objectsuch that only the normal component of the ow becomes sonic orsubsonic That is most of the ow in the combustor will remainsupersonic A schematic of the dramjet is shown in Fig 8 (fromRef 5) Here fuel is injected into the incoming airstream mixesrapidly with it and is ignited behind the centerbody of a ramjetwhich acts as a stabilizer holding the oblique wave It is similar toa conventional ramjet except for the close coupling between the re-

Fig 8 Schematic of a dramjet (from Ref 5)

Fig 9 Schematic of shock structure for scramjet engine and an ODWE(from Ref 33)

action waves and the leading shock waves resulting in a detonationwave The performance of such a device has been calculated andshown5 to be less than that of an ideal ramjet over a range of Machnumbers from 5 to 30 Issues such as the presence of inlet shocksand the ability to mix rapidly enough to obtain a detonable mixturewere not addressed in this paper5

A representation closer to scramjet engines is shown in Fig 9(from Ref 33) Here the inlet shock system is shown but the linkbetween the inlet shocks and the ODW is left unclear Once againit is assumed that fuelndashair mixing occurs rapidly enough and thatthe oblique shock wave is strong enough so that chemical reactionsoccur close behind it and are coupled to it Analysis of a conceptualvehicle using such an ODWE showed that it has better performancethan an equivalent scramjet for Mach numbers over 15 The mainadvantage arises from the reduced length of detonation engines In arelatedwork the issueof mixing inhomogeneities and their effectonthe detonation wave were addressed numerically34 There have alsobeen severalother studies examining the feasibilityand performanceof detonation wave ramjets3536 All of these concepts essentiallyrely on rapid ignition and combustion occurring behind an obliqueshock wave and the coupling between the two processes resultingin a stable structure A difference between these more recent worksand the earlier ones is the realization that in many situations shock-induced combustion does not lead to detonations However none ofthese studies seem to have led to the actual building and testing ofan ODWE

Ram AcceleratorsIf the solid body anchoring the oblique shock or detonation is

free to move the integral of pressure forces around the body canresult in a net force causing the body to accelerate or decelerateThis is essentially the concept behind the ram accelerator a propul-sion concept proposed3738 in the late 1980s In a ram accelerator

1702 KAILASANATH

Fig 10 Schematic of a ram accelerator in the detonative mode

Fig 11 Early concept of the detonative mode (from Ref 37)

a projectile that resembles the centerbody of a ramjet travels at asupersonic speed through a premixed fuelndashoxidizerndashdiluent mix-ture enclosed in a tube Because the projectile travels at supersonicspeeds oblique shock waves are formed at the nose of the projec-tile and they subsequently re ect from the side walls of the tubeand the projectile body Depending on the strength of these obliqueshocks ignition may occur after the initial shock or subsequentre ections Depending on the velocity of the projectile differentmodes of combustion and correspondingly different modes of op-eration of the system are possible When the velocity of the pro-jectile is greater than the CndashJ velocity an ODW may be stabilizedon the projectile resulting in the detonative mode of operation Aschematic of the ram accelerator operating in the detonative modeis shown in Fig 10 From the early days detonation has been dis-cussed as a means to attain very high velocities including escapevelocities37 In Ref 37 and several subsequent investigations thestructure of the detonative mode was described by the schematicdiagram shown in Fig 11 The projectile nose cone angle and thesound speed in the mixture selected were tailored to avoid combus-tion behind the initial shock wave but to provide initiation behindthe rst shock wave re ected from the tube wall Emphasis was alsoplaced on having the rst re ected shock wave strike the projec-tile body just aft of the shoulder37 If the system needed to be de-signed so precisely successful operation would indeed be dif cultin practical situations Further research has shown that the systemis far more stable and the particular scenario shown in Fig 11 ac-tually occurs just prior to system failure There were early reportsof successful attainment of superdetonative projectile velocities39

but the experimental emphasis shifted to the successful develop-ment of subdetonative modes of operation except at the Instituteof Saint-Louis (ISL) where research on the detonative mode hascontinued40 iexcl 42

Numerical simulations4344 and theoretical studies4546 based onsimpli ed analytical models of superdetonative ram acceleratorswere startedin the earlydays and have made asigni cant impactoverthe past decade These early studies con rmed the potential of ramaccelerators to accelerate ef ciently large masses to escape veloci-ties and low Earth orbit The steady-state simulations of Yungster47

highlighted the role of the projectile velocity (or Mach number)and showed that detonation could occur at different locations alongthe projectile body depending on the projectile Mach number (de- ned as the ratio of the velocity of the projectile to the speed ofsound in the premixed mixture) This work also raised the issue ofcombustion in the boundary layers along the projectile

The projectile velocity changes continuously in the ram acceler-ator and therefore steady-state simulations can at best be thoughtof as trying to characterize an instantaneous snapshot of the sys-tem The importance of the transient processes was rst highlightedin the work of Li et al48 This as well as subsequent work49 hasshown the importance of taking into account the time-dependent

Fig 12 Schematic showing the detailed structure of an oblique deto-nation and its relation to the detonation structures in ram accelerators

nature of the ow in the ram accelerator in describing phenomenasuch as unstarts and other system failures Computational studiesof the starting aspects of ram accelerators and the role of viscouseffects have also been conducted50 iexcl 52 Experimental efforts haveprimarily focused on the nondetonative mode except at ISL40 iexcl 42

Before discussing some of the experimental results and the dif cul-ties encountered in the practical development of ram accelerators inthe detonative mode two critical issues need to be discussed Oneissue is the actual structure of ODWs and the other is the stabilityof such detonation waves during transient operation

Structure and Stability of ODWsA key issue in the development of the ODWE and the ram accel-

erator is the structure and stability of ODWs Although the use ofODW was suggested as early as 195815 the stability of such waveswas implicitly assumed and questions about their detailed structurewere not considered Renewed interest in the late 1980s led to a the-oretical analysis of the supersonic ow of a combustible gas mixturepast a wedge53 Reference 53 indicated that for approach velocitiesroughly 25 or more larger than the CndashJ velocity of the reactantmixture ODWs can be stabilized over a useful range of wedge an-gles For wedge angles greater than a critical value the wave willdetach from the wedge and form an overdriven detonation wave ora more complex shock-induced combustion wave About this timethere were also numerical studies that showed that standing det-onation waves can be established on a wedge under certain owconditions5354

The early experimental efforts61314 were either focused on stabi-lizing a normal detonation wave in a duct or involved shock-inducedcombustion behind weak oblique shock waves Therefore it was notclear if ODWs could really be stabilized until some experiments inthe late 1980s on the diffraction and transmission of detonationsshowed that they could be5556

More recently the interest in ram accelerators has spawned alarge number of studies of the structure and stability of ODWs57 iexcl 61

The work of Li et al57 iexcl 59 showed that the basic structure of anoblique detonation generated by a wedge is more complex thanpreviously thought and consists of a nonreactive shock wave an in-duction region de agration waves and a detonation wave in whichthe pressure increase due to the energy release is closely coupledto the shock front This basic structure and a schematic of how itmight relate to the detonation waves in ram accelerator is shown inFig 12 These studies also showed that over a wide range of ow

KAILASANATH 1703

and mixture conditions the basic detonation structure is stable andvery resilient to disturbances in the ow In agreement with theory53

the simulations also showed the existence of a critical angle beyondwhich an attached stable detonation does not occur59 Other recentstudies have shown the same complex structure even with the inclu-sion of more complex chemical kinetic schemes6061 The essentialfeaturesof the complex ODW structure have also been con rmed byexperiments462 Thus a major conclusion from the recent researchis that the structure of an ODW can no longer be assumed to bea straight oblique shock wave followed closely by energy releaseand the complex structure can indeed be stabilizedon wedge-shapedobjects Implications of this work on propulsion applications mustbe considered

Although the schematic of Fig 12 shows how to relate the basicODW structure to that in ram accelerators the actual detonationwave structure in ram accelerators is likely to be more complexbecause of the expansion waves and multiple shock waves that aregenerated due to the geometric complexity of the system63 Theresults from a study of the detailed structure of detonations in ramaccelerators under a variety of ow conditions63 are summarizedin Fig 13 where the detonation wave structure is shown for three ow Mach numbers Recent experimental investigations have alsorevealed extremely complex detonation wave structures involvingmultiple shock waves under certain conditions64 In spite of theircomplexity all of these structures appear to be inherently relatedto the basic three-wave structure of oblique detonations shown inFig 12 However this appears to be an area of research that willcontinue for a while

a)

b)

c)

Fig 13 Detailed structure of detonations inram accelerators operatingat a) Mach 76 b) Mach 80 and c) Mach 84 (from Ref 63)

a) Case A Mach = 905

b) Case B Mach = 915

Fig 14 Visualization of the ow eld during the acceleration of theprojectile shown in Fig 5 from Mach 8 to 10 at two instants

Transient Nature of Detonation Waves in Ram AcceleratorsThe second important feature revealed by the recent studies is

the inherent transient nature of detonation waves in con gurationssuch as the ram accelerator When the concept of ram acceleratorsin the detonative mode was rst proposed it was thought that theleading shock wave (from the nose of the projectile) should re ectfrom the tube walls such that it impinges exactly at the shoulderof the projectile This will maximize the thrust and hence accel-eration However if the system were designed so precisely for onecondition its off-design performance could be unacceptable Fur-thermore simulations have shown that as the projectile acceleratesthe detonation wave structure tends to also move49 To illustrate thedynamics of detonations in the ram accelerator two photographsof the ow eld from a simulation (where the projectile acceleratedfrom Mach 8 to 10) are shown in Fig 14 In case A the projectileMach number is 905 and in case B it is 915 For both cases themain illustrations show the concentration of water vapor with thetwo inserts showing the pressure and temperature in a selected re-gion of the ow eld At the instant shown in Fig 14a (case A) thepressure and temperature behind the rst re ected shock wave arenot very high and combustion and energy release (as indicated bythe water vapor concentration) takes place only behind the secondre ected shock wave That is the detonation can be said to occurbehind the second re ected shock wave By the time the projectilehas accelerated to Mach 915 the leading shock wave has strength-ened raising the pressure and temperature behind the rst re ectedshock wave and causing the detonation to occur there

The behavior just discussedis seenthroughout the operationof theram accelerator in the detonative mode That is detonation waves inram accelerators are dynamic and move from shock wave to shockwave as the projectile acceleratesEventually the detonation wouldoccur just after the rst shock wave and the whole system will ceaseto operate because there will not be any positive thrust Furthermorethe detonation wave will run in front of the projectile creating a sit-uation similar to the classical unstart phenomena in ramjets Suchunstarts and detonations occurring after a different number of re- ections have also been noted in experiments42 Hence the shape ofthe projectile has to be chosen carefully so that the detonation canmove over the projectile body and still produce positive thrust overthe regime of interest A practical problem that has hindered thedevelopment of ram accelerators operating in the detonative modeis the change in the shape of the projectile due to material erosionburning and failure This is currently being overcome by the use ofsteel cowlings42

1704 KAILASANATH

PDEsThe other major current propulsion application involving detona-

tion waves is the PDE The basic concept behind a PDE is simpleDetonation is initiated repeatedly at either the closed or the openend of a detonation chamber that is lled with a premixed fuelndashairmixture Figure 15 (based on Ref 65) illustrates the concept forthe case when the detonation wave is initiated at the closed endAs shown in this schematic a one-dimensional planar detonation isinitiated near the closed end of the detonation chamber and travelsat the CndashJ velocity toward the open end A set of rarefaction wavestrail the detonation wave to reduce the velocity to zero at the closedend of the tube When the planar detonation wave exits the chamberanother set of expansion waves is generated that travels toward theclosed end These waves evacuate the burned detonation productsresulting in a cool empty chamber that is ready to be lled with afresh fuelndashair mixture The entire cycle repeats resulting in a pe-riodic high-pressure zone near the closed end of the chamber Theintegrated effectof this high pressure over the closed end (represent-ing a thrust wall) produces thrust This is of course a very idealizedpicture of the system Before discussing progress made in the re-cent years in developing a PDE a brief review of the background isprovided

The early work on the use of intermittent detonations has alreadybeen discussed In the late 1980s this concept of using intermittentor pulsed detonations was reexamined experimentally at the NavalPostgraduate School66 An ethylenendashoxygen detonation wave in asmall-diameter tube was used as a predetonator to initiate detona-tions in a larger tube containing an ethylenendashair mixture Periodicfuel injection within the naturally aspirated tube resulted in an in-termittent frequency of 25 Hz Speci c impulse estimated usingthe pressure time history and the amount of fuel consumed rangedfrom 1000 to 1400 s The velocities of the observed waves (lessthan 1 kms) are signi cantly below the CndashJ detonation wave ve-locities for the reported mixtures indicating that a fully developeddetonation wave was not formed

The rst purely computational study of the PDE reported in theopen literature appears to be the work of Cambier and Adelman67

The system simulated consisted of a 50-cm-long main tube attachedto a 43-cm-long diverging nozzle Quasi-one-dimensional simula-tions using the Euler equations for unsteady ow were carried outwith a total variation diminishing scheme and multistep nite ratekinetics A detonation was initiated at the closed end of a tube lledwith a premixed stoichiometric hydrogenndashair mixture at 3 atmOverall performance calculated by integrating the instantaneousthrust and the fuel ow rates gave a speci c impulse Isp of about6500 s and a range of operating frequencies up to 667 Hz Issues ofignition energies and transition to detonation were not consideredand the dif culties with attaining rapid mixing of fuel and air wasavoided by using a premixed fuelndashair mixture These assumptionsare still typical of those used in more recent simulations

The rst look at the interactions between the ow inside the cham-ber and that outside appears to be the work of Eidelman et al7

who simulated a cylindrical detonation chamber (about 15-cm diam

Fig 15 Schematic of an idealized PDE showing various stages duringone cycle of operation (adapted from Ref 65)

Fig 16 Generic PDE device used for somecomputational studieswheredetonation is initiated near the open end (from Ref 69)

and 15 cm long) with a small converging nozzle In their two-dimensional simulations a self-similarsolution for a planar detona-tion was imposed near the open end and made to travel toward theclosed end The re ection of the wave from the closed end createda higher pressure at the wall than during initiation near the closedend Hence this mode of initiation was considered to be better Af-ter the detonation left the chamber the pressure in the chamber wasobserved to fall below the ambient resulting in ingestion of outsideair In subsequent works86869 the same numerical techniques andmodels were used to investigate a slightly different con gurationwhere the nozzle at the aft end was removed and an inlet was in-troduced near the closed end This con guration has become oneof the popular systems for two-dimensional (and axisymmetric) nu-merical simulations Again detonation was assumed to be initiatedat the open end and traveled toward the closed end In this con gu-ration when the pressure falls below atmospheric in the detonationchamber air is ingested through the inlets in the front even duringstatic operations Thus a valveless self-aspiratingoperation is pos-sible These simulations also introduced the discussion of open-endinitiation vs closed-end initiation and their relative advantages Thecurrent status of this issue will be discussed later

Many of the early efforts on intermittent or PDEs have been re-viewed in Ref 7 and 8 where a link is also made between thesedetonation engines and pulse jet engines (which do not use detona-tive combustion) In addition Ref 8 provides a detailed summaryof the numerical investigations conducted by the authorsrsquo group ona generic PDE device shown in Fig 16 A characteristic feature ofthis device is that the detonation wave is initiated at the open endand travels toward a thrust wall at the closed end An advantage ofthis approach is that it is valveless and self-aspirating Air is en-trained into the chamber through inlets near the closed end whenthe pressure falls below atmospheric pressure during part of the cy-cle Unsteady Euler simulations were used to study the operationof the device from static (M = 0) to supersonic (M = 2) conditionsAs stated in Ref 8 a proof of principle experimental demonstra-tion of the PDE mode of operation where detonation is initiated atthe open end still needed to be done Furthermore issues such asdetonation initiation and structure and fuelndashair injection and mixingalso needed to be studied

Bussing and Pappas65 provided a detailed description of thebasic operation of an ideal zed PDE They also reported someone-dimensional studies of PDEs burning hydrogenndashoxygen andhydrogenndashair mixtures The detonation wave was initiated at theclosedend in this engine using a high-temperature and high-pressureregion An advantage of this design is the con nement provided atthe closed end which should enable a more rapid initiation andestablishment of a detonation Both airbreathing and rocket modeof operation of this valved design was discussed Lynch et al70

presented a computational study of an axisymmetric PDE with astraight inlet very similar to that of Eidelman and Grossmann8 andEidelman et al69 Various inlet lengths from 1 to 12 cm and ightMach numbers of 08 and 20 were studied The results were similarto those reported earlier except that some additional details of thedetonation wave front and the air induction system were resolved inthese studies In a related work Lynch and Edelman71 showed thatreductions in scavenging time (and hence overall cycle time) canbe achieved by suitably shaping the inlet

Bussing et al72 compared open-end initiation and closed-end ini-tiation in a simple tube PDE geometry Equivalent thrust productionand fuel ef ciency were observed for the two types of initiation

KAILASANATH 1705

Although a higher head-end pressure was attained during open-endinitiation the pressure continuously dropped for that case unlike theclosed-end initiation where a constant pressure region occurred fora while before reduction of the pressure to the ambient value Morerecently an experimental study has con rmed the differences in theobserved pressure pro les and it has also been concluded that themaximum level of impulse attained is independent of the locationof the detonation initiator73

In the past few years there has been a substantial growth in theappearance of papers related to the PDE and complete sessionshave been devoted to the topic at annual AIAAASMESAEASEEjoint propulsion conferences Because of space considerations onlyselected papers from these years are brie y discussed here Sterlinget al74 conducted one-dimensional numerical investigations of aself-aspiratingPDE operating for multiple cyclesA key observationfrom their studies was that only a portion of the detonation tubecan be lled with fresh charge under self-aspirating operation Inspite of this limitation they reported a speci c impulse of 5151 sfor a hydrogenndashair system However they concluded that the idealperformance of such an engine is ldquonear those of other hydrogen-fueledair breathing enginesrdquo74

Several basic shock tube experiments related to the developmentof a hydrogen-fueled PDE were described by Hinkey et al75 Theseinclude the measurement of detonation wave velocities and de a-gration to detonation transition (DDT) lengths for a range of equiv-alence ratios As expected the DDT lengths were found to be toolarge for practical applications even in hydrogenndashoxygen mixturesTherefore traditional techniques such as the use of a Schelkin spiral(see Ref 76) were tried to reduce the transition length A factor of2ndash4 reduction was observed over the range of equivalence ratios in-vestigated This work75 highlights the dif culties involved in obtain-ing a fully developed detonation in a short distance (as assumed inthe conceptual studies and numerical simulations discussed earlier)and questions if the detonations reported in previous experimentalinvestigations161766 were really fully transitioned from de agra-tions The speci c impulses measured in their experiments (about240 s for hydrogenndashoxygen and 1200 s for hydrogenndashair) were alsostated to be in agreement with their analysis

An interesting PDE concept presented by Bussing77 was that of arotary-valved multiple-pulsed detonation engine Here several det-onation chambers are coupled to an air inlet and fuel source usinga rotary valve as suggested in one of the early papers of Nichollset al16 The rotary valves allow the lling of some detonation cham-bers while others are detonating or exhausting Other papers includepresentation of an overall PDE performance model78 and the exper-imental characterization of the detonation properties of some fuelsfor use in PDEs79

The importance of adequately mixing the fuel and oxidizer washighlighted by the experimental investigations of Stanley et al80

who obtained very low sub-CndashJ velocities when injecting the fueland oxidizer at different times and not taking extra effort to en-sure that they were well mixed The use of turbulence producingdevices appeared to improve signi cantly the mixing and the at-tainment of higher velocities but also resulted in signi cant thrustlosses This study also showed that raising the initial pressure wasbene cial The computational studies of Eidelman et al81 showedthat the PDE engine could operate even for a range of transitionaldetonation regimes that produced nonplanar or not fully developeddetonations Aarnio et al82 discussed two failure modes DDT tran-sition failure and premature ignition In both cases the system con-tinued to operate though there would be loss of thrust In Ref 82detailed discussions of the pressure and thrust histories during a5-Hz 20-cycle operation are also provided (Fig 17) The cycle tocycle variations of peak pressure have been attributed to the lowsampling frequency used82 Both a direct measurement as well asthe integration of the pressure histories were used to estimate the av-erage Isp from hydrogenndashair systems to be between 1116 and 1333 sThe effect of the predetonator volume on the reduction of speci cimpulse was also shown to be a critical issue The effects of inletsand nozzles were not considered in this study

A conceptual design of a multitube PDE with a single air inletduct was discussed by Pegg et al83 Time-dependent computational

Fig 17 Pressure history at one location from a PDE operating at 5 Hzfor 20 cycles (from Ref 82)

uid dynamics analysis indicated that the inlet isolaterdiffusercon-cept would work and did not allow enough time for the formationof destabilizing hammer shocks Performance analysis includingcomponent ef ciencies reported in the paper indicate that opera-tional frequencies of the order of 75ndash100 Hz are required for theMach 12ndash3 ights considered Experimental data indicating mul-ticycle operation at 100 Hz with a rich ethylenendashoxygen mixturewere presented by Sterling et al84 DDT enhancement devices andpredetonators were used in the experiments In spite of this det-onation waves did not occur all of the time and when detonationwaves failed the tube became very hot and testing was suspendedTherefore the maximum testing time was reported to be 05 s Inspite of this limitation the two studies8384 taken together high-light the possibility of PDEs burning hydrocarbon fuels for realisticmissions

Five different nozzle shapes and their effect on performance werestudied computationally by Cambier and Tegner85 Their results in-dicate that the presence of a nozzle can affect the performance ofthe PDE by increasing the thrust delivery during the ignition phaseThe bell-shaped nozzles appeared to give higher performance thanshapes with positive curvature Their results also showed that noz-zles also affected the ow dynamics and hence the timing of thevarious phases of the engine cycle In an experimental study of aconical exhaust nozzle the authors of Ref 86 noted improved per-formance (based on higher velocities which were still sub-CndashJ) butdid not see any effect on the blowdown process More recently87

the effectof various nozzle shapes including converging divergingand straight have been reexamined computationally The converg-ing sections of nozzles introduced shock wave re ections whereasdiverging sections generated a negative thrust for a portion of thecycle due to overexpansion In spite of these limitations the overallconclusion of these studies was that ldquonozzles can drastically in-crease ef ciency of the PDEsrdquo87 There is also some experimentalevidence suggesting that propulsive performances can be increasedby the addition of a nozzle73 However factors such as the effect ofthe nozzles on detonation transmission and the detailed dynamicsof the ow have not yet been elucidated This appears to be an areathat needs further investigation

Other recent efforts have focused on demonstrating the opera-tion of single- and multitube combustors coupled to an inlet using arotary-valvemechanism88 The valve servesto both meter the air owinto the combustor and to isolate the inlet from the high pressuresproduced during the detonation cycle Utilizing multiple combus-tors that ll and detonate out of phase allows the continual use of thein owing air Firing rates of up to 12 Hz per combustor were demon-strated for a hydrogen-fueled system Several other issues related tothe practical development of PDEs have been discussed by Bussinget al89 According to them three of the more important engineer-ing issues are the development of a valve-injection subsystem anignition system and a low-volume predetonator

The use of PDEs for rockets has been revived recently90 Therocket mode of operation is very similar to the description providedearlier of the airbreathing engine with ignition at the closed endexcept that the oxidizer also needs to be injected into the system pe-riodically An advantage of PDEs for rockets would be their higherpower density thus enabling the development of more compact

1706 KAILASANATH

rockets Reference 90 shows the pressure traces from a pulsed det-onation rocket engine operating at 145 Hz on a hydrogenndashoxygenmixture

More recently a two-tube rotary-valved PDE with ight-sizecomponents was operated at 40 Hz per combustor for 30 s withan ethylenendashair mixture91 In addition throttling of the device al-lowed operation at three thrust levels during a 10-s demonstrationexperiment Contrary to some earlier estimates the thermal loadswere signi cant and it was not possible to operate for more than3 s without water cooling It is not clear if this was because ofperiodic detonation failures and consequent de agrative modes ofcombustion as reported in some earlier studies

The work reported in the open literature and discussed earlier hasfocused on gaseous fuels However for volume-limited propulsionapplications PDEs operating on liquid fuels need to be demon-strated In a recent study Brophy et al92 reported on experimentsusing JP-10oxygen and JP-10air aerosols The fuelndashoxygen mix-ture was successfully detonated to obtain an engine operating at5 Hz but the tests with the fuelndashair mixture were not successful

Computational studies continue to make progress in the study ofthe PDE8793 iexcl 97 The effect of nozzle shapes and incomplete transi-tion to CndashJ detonations have been studied by Eidelman and Yang87

They nd that the cycle ef ciency will be virtually the same whethera CndashJ detonation wave is initiated instantaneously or if the CndashJ valueis attainedonly at the end of the detonation chamber Primarilybasedon this study87 they come to an interesting conclusion that ldquoinitia-tion energies that are required for PDE operations can be small andcomparable to energies required for initiation of (other) combustionprocessesrdquo However this is yet to be substantiated experimentallySekar et al93 reexamine open- and closed-end initiation and thebasic self-aspirating con guration of Eidelman8 using more recentcomputational tools Injection and mixing issues that have gener-ally been ignored in previous numerical studies are also beginningto be addressed94 Both front-wall and lateral-wall injections werestudied and the time required for fuel injection was estimated Itis suggested that it may be possible to ll only a portion of thechamber during high-frequency operations Recent computationalstudies also suggest that preconditioning the fuelndashair mixture us-ing a shock wave can signi cantly reduce the DDT distance95 Theperformance estimates mentioned throughout the discussion of thePDE show signi cant variation A systematic review of various per-formance estimates and possible reasons for the observed variationhave been presented9697 These numerical studies suggest that thetime history of the back pressure is a crucial parameter affecting theestimated performance

Other Detonation EnginesThere have also been several other interesting conceptual studies

on the applications of detonations to engines Cambier et al98 haveproposed a pulsed-detonation wave augmentation device as part ofa hybrid engine for a single-stage-to-orbit airbreathing hypersonicvehicle Here the PDE is used for both thrust generation as well asfor mixingcombustion augmentation Two speci c con gurationswere investigated numerically

The concept of a detonation internal combustion engine was pro-posed by Loth and Loth99 and Loth100 Here a conventional pistoncylinder con guration is modi ed to incorporate a separate deto-nation chamber which is isolated by a valve from the compressionchamber during fuel injection and discharges tangentially into a vor-texchamber formedby the piston and cylinder at top deadcenterTherapid detonative combustion is followed by rapid dilution throughmixing with air in the vortex chamber to reduce the formation ofNOx and unburned hydrocarbons and to achieve an overall leanmixture The vortex chamber was also designed to store a portion ofthe detonation waversquos kinetic energy Overall thermal ef cienciessomewhat better than constant volume combustion were postulatedunder ideal conditions

ConclusionsDetonation waves have been explored extensively for propulsion

applications because of their inherent theoretical advantage overde agrative combustion However practical developments to date

have been of idealized systems or laboratory-scale devices or bothThe basic ideas behind most of the current systems being investi-gated have been known for a while but still many of the details needto be understood Advances in computations and experimental di-agnostics appear to be poised to make practical propulsion devicesbased on detonation waves a reality

Two major systems currently under development are the ram ac-celerator and the PDE In both cases the theoretical and computa-tional studies have been far more encouraging than the experimentsHowever in both cases actual demonstrations of devices workingon gaseous fuels have taken place In the case of the ram accelera-tor fuelndashoxygenndashdiluent mixtures have been used and the primaryproblem appears to be heat transfer and related material failuresThe use of gaseous fuels is not a major issue here but new materialsfor the projectiles need to be explored

In the case of the PDE most of the successful operations havebeen with fuelndashoxygen mixtures There still has not been a directcomparison between detailed experimental data and related theoret-ical and computational results Furthermore most of the PDE andrelated work to date have focused on gaseous mixtures PDEs op-erating on multiphase mixtures need to be emphasized because itwould be more practical to use liquid fuels for most of the propul-sion applicationsproposed for these systemsThere are several othergeneral issues that need to be resolved Clearly the evacuation ofexhaust gases is important both from the point of view of controllingthe cycle time and for preventing autoignition and mixing with freshgases The heat transfer and noise from PDEs have also not beenreported in any detail Other issues include inletndashcombustor andcombustorndashnozzle interactions and scalability of the entire device

AcknowledgmentsThis work has been sponsored by the Of ce of Naval Research

through the Mechanics and Energy Conversion Division and theUS Naval Research Laboratory

References1Fickett W and Davis W C Detonation Univ of California Press

Berkeley CA 19792Hoffmann N ldquoReaction Propulsion by Intermittent Detonative Com-

bustionrdquo German Ministry of Supply AI152365 Volkenrode Translation1940 (cited in Ref 16)

3Roy M ldquoPropulsion par Statoreacteur a Detonationrdquo Comptes-Rendusde lrsquoAcademie des Sciences Vol 222 1946 pp 31 32

4Dabora E K and Broda J C ldquoStanding Normal Detonations andOblique Detonations for Propulsionrdquo AIAA Paper 93-2325 July 1993

5Dabora E K ldquoStatus of Gaseous Detonation Waves and Their Rolein Propulsionrdquo Fall Technical Meeting of the Eastern States Section ofthe Combustion Institute Combustion Inst Pittsburgh PA 1994 pp 11ndash

186Rubins P M and Bauer R C ldquoReview of Shock-Induced Supersonic

Combustion Research and Hypersonic Applicationsrdquo Journal of Propulsionand Power Vol 10 No 5 1994 pp 593ndash601

7Eidelman S Grossmann W and Lottati I ldquoReview of PropulsionApplications and Numerical Simulations of the Pulsed Detonation EngineConceptrdquo Journal of Propulsion and Power Vol 7 No 6 1991 pp 857ndash

865 also AIAA Paper 89-2446 July 19898Eidelman S and Grossmann W ldquoPulsed Detonation Engine Experi-

mental and Theoretical Reviewrdquo AIAA Paper 92-3168 July 19929Reingold L ldquoRecherches sur les combustions permanentes apportees

aux foyers a circulation interne supersoniquerdquo ONERA TN2 Dept of En-ergy and Propulsion Study 728-E Paris March 1950

10Bitondo D and Bollay W ldquoPreliminary Performance Analysis ofthe Pulse-Detonation-Jet Engine Systemrdquo Aerophysics Development CorpRept ADC-102-1 April 1952

11Gross R A ldquoExploratory Study of Combustion in Supersonic FlowrdquoUS Air Force Of ce of Scienti c Research TN 59-587 Washington DCJune 1959

12Gross R A and Chinitz W ldquoA Study of Supersonic CombustionrdquoJournal of Aerospace Science Vol 27 No 7 1960 pp 517ndash525

13Nicholls J A Dabora E K and Gealer R L ldquoStudies in Con-nection with Stabilized Gaseous Detonation Wavesrdquo Seventh Symposium(International) on Combustion Combustion Inst Pittsburgh PA 1959 pp766ndash772

14Nicholls J A and Dabora E K ldquoRecent Results on Standing Detona-tion Wavesrdquo Eighth Symposium (International) on Combustion CombustionInst Pittsburgh PA 1962 pp 644ndash655

KAILASANATH 1707

15Dunlap R Brehm R L and Nicholls J A ldquoA Preliminary Studyof the Application of Steady-State Detonative Combustion to a ReactionEnginerdquo Jet Propulsion Vol 28 No 7 1958 pp 451ndash456

16Nicholls J A Wilkinson H R and Morrison R B ldquoIntermittentDetonation as a Thrust-Producing Mechanismrdquo Jet Propulsion Vol 27No 5 1957 pp 534ndash541

17Krzycki L J ldquoPerformance Characteristics of an Intermittent Detona-tion Devicerdquo US Naval Ordnance Test Station US Naval Weapons Rept7655 China Lake CA 1962

18Nicholls J A Cullen R E and Ragland K W ldquoFeasibility Studiesof a Rotating Detonation Wave Rocket Motorrdquo Journal of Spacecraft andRockets Vol 3 No 6 1966 pp 893ndash898

19Adamson T C and Olsson G R ldquoPerformance Analysis of a RotatingDetonation Wave Rocket Enginerdquo Astronautica Acta Vol 13 No 4 1967pp 405ndash415

20Shen P I and Adamson T C ldquoTheoretical Analysis of a RotatingTwo-Phase Detonation in Liquid Rocket Motorsrdquo Astronautica Acta Vol17 No 4 1972 pp 715ndash728

21Clayton R M and Rogero R S ldquoExperimental Measurements on aRotating Detonation-Like Wave Observed During Liquid Rocket ResonantCombustionrdquo Jet Propulsion Lab TR 32-788 California Inst of Technol-ogy Pasadena CA Aug 1965

22Back L H ldquoApplication of Blast Wave Theory to Explosive Propul-sionrdquo Acta Astronautica Vol 2 1975 pp 391ndash407

23Varsi G Back L H and Kim K ldquoBlast Wave in a Nozzle for Propul-sion Applicationsrdquo Acta Astronautica Vol 3 1976 pp 141ndash156

24Kim K Varsi G and Back LH ldquoBlast Wave Analysis for DetonationPropulsionrdquo AIAA Journal Vol 10 No 10 1977 pp 1500ndash1502

25Back L H Dowler W L and Varsi G ldquoDetonation PropulsionExperiments and Theoryrdquo AIAA Journal Vol 21 No 10 1983 pp 1418ndash

142726Kantrowitz A ldquoPropulsion to Orbit by Ground Based Lasersrdquo Astro-

nautics and Aeronautics Vol 10 No 5 1972 p 7427Hyde R A ldquoOne-D Modelling of a Two-Pulse LSD Thrusterrdquo Pro-

ceedings of the 1986 SDIODARPA Workshop on Laser Propulsion editedby J T Kare Lawrence Livermore National Labs 1987 pp 79ndash88

28Emery M H Kailasanath K Oran E S and Gardner J H ldquoCompu-tational Studies of Laser Supported Detonationsrdquo Proceedings of the 1987SDIO Workshop on Laser Propulsion edited by J T Kare Lawrence Liv-ermore National Labs 1990 pp 43ndash55

29Emery M H Kailasanath K and Oran E S ldquoNumerical Simu-lations of Laser Supported Detonationsrdquo 1988 SDIO Workshop on LaserPropulsion Lawrence Livermore National Labs April 1988

30Carrier G F Fendell F McGregor D Cook S and Vazirani MldquoLaser-Initiated Conical Detonation Wave for Supersonic CombustionrdquoAIAA Paper 91-0578 Jan 1991

31Fendell F E Mitchell J McGregor R Magiawala K and Shef eldM ldquoLaser-Initiated Conical Detonation Wave for Supersonic CombustionIIrdquo AIAA Paper 92-0088 Jan 1992

32Carrier G F Fendell F E and Chou M S ldquoLaser-Initiated ConicalDetonation Wave for Supersonic Combustion IIIrdquo AIAA Paper 92-3247July 1992

33Menees G P Adelman H G Cambier J-L and Bowles J V ldquoWaveCombustors for Trans-Atmospheric Vehiclesrdquo Journal of Propulsion andPower Vol 8 No 3 1992 pp 709ndash713

34Cambier J L Adelman H and Menees G ldquoNumerical Simulationsof an Oblique Detonation Wave Enginerdquo AIAA Paper 88-0063 1988

35Atamanchuk T and Sislian J ldquoOn- and Off-design Performance Anal-ysis of Hypersonic Detonation Wave Ramjetsrdquo AIAA Paper 90-2473 1990

36Kuznetsov M M Neyland V J and Sayapin G N ldquoEf ciency In-vestigation of Scramjet with Detonation and Shockless Supersonic Combus-tionrdquo Utchenye Zapisky TsAGI Vol 23 No 2 1992 pp 30ndash37

37Bruckner A P and Hertzberg A ldquoRam Accelerator Direct LaunchSystem for Space Cargordquo International Astronautical Federation Paper 87-211 1987

38Hertzberg A Bruckner A P and Bogdanoff D W ldquoRam Accel-erator A New Chemical Method for Accelerating Projectiles to UltrahighVelocitiesrdquo AIAA Journal Vol 26 No 2 1988 pp 195ndash203

39Kull A Burnham E Knowlen C Bruckner A P and HertzbergA ldquoExperimental Studies of the Superdetonative Ram Accelerator ModesrdquoAIAA Paper 89-2632 1989

40Patz G Seiler F Smeets G and Srulijes J ldquoStatus of ISLrsquos RA-MAC30 with Fin Guided Projectiles Accelerated in a Smooth Borerdquo Pro-ceedings of the Second International Workshop on Ram Accelerators Univof Washington Seattle WA 1995

41Seiler F Patz G Smeets G and Srulijes J ldquoGasdynamic Limitsof Ignition and Combustion of a Gas Mixture in ISLrsquos RAMAC30 ScramAcceleratorrdquo Proceedings of the Second International Workshop on RamAccelerators Univ of Washington Seattle WA 1995

42Seiler F Patz G Smeets G and Srulijes J ldquoIn uence of ProjectileMaterial and Gas Composition on Superdetonative Combustion in ISLrsquosRAMAC30rdquo AIAA Paper 98-3445 July 1998

43Brackett D C and Bogdanoff D W ldquoComputational Investigation ofOblique Detonation Ramjet-in-Tube Conceptsrdquo Journal of Propulsion andPower Vol 5 No 3 1989 pp 276ndash281

44Yungster S Eberhardt S and Bruckner A P ldquoNumerical Simula-tion of Shock-Induced Combustion Generated by High-Speed Projectiles inDetonable Gas Mixturesrdquo AIAA Paper 89-0673 1989

45Rom J and Kivity Y ldquoAccelerating Projectiles Up to 12 kms Utiliz-ing the Continuous Detonation Propulsion Methodrdquo AIAA Paper 88-29691988

46Humphrey J W ldquoParametric Study of an ODW Scramaccelerator forHypersonic Test Facilitiesrdquo AIAA Paper 90-2470 1990

47Yungster S ldquoNavierndashStokes Simulation of the Supersonic CombustionFlow eld in a Ram Acceleratorrdquo AIAA Paper 91-1916 1991

48Li C Kailasanath K Oran E S Boris J P Paper and LandsbergA M ldquoNumerical Simulations of Transient Flows in Ram AcceleratorsrdquoAIAA Paper 93-1916 1993

49Li C Kailasanath K Oran E S Landsberg A M and Boris J PldquoDynamics of Oblique Detonations in Ram Acceleratorsrdquo Shock Waves Vol5 June 1995 pp 97ndash101

50Yungster S and Radhakrishnan K ldquoComputational Study of FlowEstablishment in a Ram Acceleratorrdquo AIAA Paper 98-2489 July 1995

51Choi J Y Jeung I-S and Yoon Y ldquoNumerical Study of ScramAccelerator Starting Characteristicsrdquo AIAA Journal Vol 36 No 6 1998pp 1029ndash1038

52Bezgin L Ganzhelo A Gouskov O and Kopchenov V ldquoSome Nu-merical Investigation Results of Shock-Induced Combustionrdquo Proceedingsof the AIAA 8th International Space Planes and Hypersonic Systems andTechnologies Conference AIAA Reston VA 1998 pp 72ndash82

53Pratt D W Humphrey J W and Glenn D E ldquoMorphology of Stand-ing Oblique Detonation Wavesrdquo Journal of Propulsion and Power Vol 7No 5 1991

54Fujiwara T Matsuo A and Nomoto H ldquoTwo-Dimensional Detona-tion Supported by a Blunt Body or a Wedgerdquo AIAA Paper 88-0089 Jan1988

55Fan B C Sichel M and Kauffman C W ldquoAnalysis of ObliqueShock-Detonation Wave Interactions in the Supersonic Flow of a Com-bustible Mediumrdquo AIAA Paper 88-0441 Jan 1988

56Dabora E K Desbordes D Guerraud C and Wagner H GldquoOblique Detonations at Hypersonic Velocitiesrdquo Dynamics of Detonationsand Explosions Detonations Vol 133 Progress in Astronautics and Aero-nautics AIAA Washington DC 1990 pp 187ndash201

57Li C Kailasanath K and Oran E S ldquoStability of Oblique Detona-tions in Ram Acceleratorsrdquo AIAA Paper 92-0089 Jan 1992

58Li C Kailasanath K and Oran E S ldquoStructure of Reaction WavesBehind Oblique Shocksrdquo Dynamic Aspects of Detonation Vol 153 Progressin Astronautics and Aeronautics AIAA Washington DC 1993 pp 231ndash

24059Li C Kailasanath K and Oran E S ldquoDetonation Structures Behind

Oblique Shocksrdquo Physics of Fluids Vol 6 April 1994 pp 1600ndash161160Da Silva L F and Deshaies B ldquoNumerical Study of Some Ignition

Regimes of Combustible Supersonic Flows over a Wedgerdquo Proceedings ofSecond International Workshop on Ram Accelerators Paper 15 Univ ofWashington Seattle WA 1995

61Thaker A A and Chelliah H K ldquoNumerical Prediction of ObliqueDetonation Wave Structures Using Detailed and Reduced Reaction Mecha-nismsrdquo Combustion Theory and Modelling Vol 1 No 4 1997 pp 347ndash376

62Desbordes D Guerraud C and Hamada L ldquoSupersonic H2-AirCombustions Behind Oblique Shock Wavesrdquo 14th International Colloquiumon Dynamics of Explosions and Reactive Systems Paper D28 Univ ofCoimbra Portugal Aug 1993

63Li C Kailasanath K and Oran E S ldquoDetonation Structures Gener-ated by Multiple Shocks on Ram-Accelerator Projectilesrdquo Combustion andFlame Vol 108 1997 pp 173ndash186

64Viguier C and Desbordes D ldquoConditions of Onset and Stabilizationof Oblique Detonation Wavesrdquo Advances in Experimentation and Compu-tation of Detonations edited by G D Roy S H Frolov K Kailasanathand N Smirnov ENAS Moscow 1998 pp 40 41

65Bussing Tand Pappas G ldquoIntroduction toPulseDetonation EnginesrdquoAIAA Paper 94-0263 1994

66Helman D Shreeve R P and Eidelman S ldquoDetonation Pulse En-ginerdquo AIAA Paper 86-1683 June 1986

67Cambier J L and Adelman H G ldquoPreliminary Numerical Simu-lations of a Pulsed Detonation Wave Enginerdquo AIAA Paper 88-2960 July1988

68Eidelman S Grossmann W and Lottati I ldquoAir-Breathing PulsedDetonationEngine Concept A Numerical Studyrdquo AIAA Paper 90-2420 July1990

69Eidelman S Grossmann W and Lottati I ldquoComputational Analysisof Pulsed Detonation Engines and Applicationsrdquo AIAA Paper 90-0460 Jan1990

KAILASANATH 1699

Early ResearchAlthough basic studies and applications of detonation have been

undertaken for a very long time speci c references23 to propulsionappear in the literatureonly in the 1940s Even at this early time bothstanding (or stabilized) and unsteady (intermittent)detonations wereexplored In the work of Hoffmann2 both gaseous (acetylene) andliquid (benzene) hydrocarbon fuels were employed with oxygenIntermittent detonation appears to have been achieved but attemptsto determine an optimum cycle frequency were less successful Thedevelopment of the concept of pulsed detonation engines (PDEs)has been traced back to this pioneering work in a number of papersThe proposals of Roy3 inspired further work in France on the de-sign of systems to stabilize combustion in supersonic ows (egRef 9) Soon work was also begun in the United States Bitondoand Bollay10 conducted an analytical study that indicated that a

Fig 1 Comparison of idealized thermodynamic cycles for constantpressure constant volume and detonation modes of combustion

Fig 2 Schematic of a standing detonation wave at the exit of a nozzle(from Ref 14)

Fig 3 Schematic of an experimental setup for multicycle detonation studies (from Ref 16)

pulsating detonation engine can be helpful for helicopter propul-sion Gross11 Gross and Chinitz12 Nicholls et al13 and Nichollsand Dabora14 studied means to stabilize detonation waves for ap-plications to hypersonic ramjet propulsion The possibility of usingoblique detonations was also introduced15 More recent extension ofthese ideas will be discussed in the sections on ldquoOblique DetonationWave Enginesrdquo (ODWEs) and ldquoRam Acceleratorsrdquo

Stabilized Normal DetonationsFirst we brie y look at some of the early experiments to stabi-

lize detonation waves A schematic of an experiment is shown inFig 2 (from Ref 14) Cold hydrogen gas injected at the throat ofa convergentndashdivergent nozzle mixes with the high-pressure high-temperature air owing through the nozzle Because of the shortresidence time and rapid drop in temperature combustion does notoccur within the nozzle The nozzle is operated highly underex-panded so that further expansion occurs outside resulting in a com-plex system of shock waves a key feature of which is the Machdisk a nearly normal shock The conditions behind this Mach diskare such that ignition and energy release occur just behind it If theenergy release is closely coupled to the shock wave a detonationwill be established Whether or not a detonation is established willdepend on the induction time behind the Mach disk which in turnwill depend on the pressure temperature and mixture compositionApplication of such steady-state detonations to propulsion were ex-plored and performance comparable to conventional ramjets werereported for appreciably higher ight Mach numbers15

Intermittent DetonationsConcurrent with their work on stabilized detonation waves

Nicholls et al16 also explored the concept of intermittent (or pulsed)detonation waves for propulsion applications Both single-cycle andmulticycle operations with hydrogen and acetylene as fuels andoxygen and air as oxidizers were demonstrated The basic setupwas a simple detonation tube open at one end with coannular fueland oxidizer injection at the closed end as shown in Fig 3 (fromRef 16) Thrust fuel ow air ow and temperature measurementswere made over a range of operating conditions When the sparkplug was located 2 in (508 cm) or 5 in (127 cm) from the end ofthe mixing plane spasmodic ring was observed suggesting prob-lems with fuelndashair mixing However with the spark plug located 10in (254 cm) downstream periodic detonations for a range of mix-tures were reported For a hydrogenndashair mixture a speci c impulseof 2100 s was attained along with a cycle frequency of 35 Hz Theyalso presented a very simpli ed theoretical analysis that gave over-all results very much in agreement with their experimental data onhydrogen but less with the case of acetylene16 However they real-ized that the agreement was partly fortuitous because the measuredthrust-time history was signi cantly different from the theoreticalresult (Fig 4 from Ref 16) Note that they also attempted initiationfrom the open end but were not successful

A setup similar to that of Nicholls et al16 was constructed byKrzycki17 who used automotive spark plugs for ignition He demon-strated operation at 60 Hz with propanendashair mixtures but there issome doubt whether the device was operating in the detonative mode

1700 KAILASANATH

Fig 4 Comparison of theoretical and experimental results on timehistories of detonation tube acceleration for a 50 acetylenendashoxygendetonation (from Ref 16)

Fig 5 Possible con guration for a rotating detonation wave engine(from Ref 19)

or merely as a pulse jet engine due to the low initiation energies em-ployed His overall conclusion was that although thrust was possiblefrom such a device practical applications were not promising

Rotating DetonationAnother interesting concept that was explored was that of a rotat-

ing detonation wave rocket motor18 iexcl 20 Here gaseous fuel (hydrogenor methane) and oxygen were continuously injected into an annularcombustion chamber and ignited using spark plugs A unidirec-tional detonation wave was created after transition from a de agra-tion wave and the exhaust gases were expelled through an annularnozzle Such a system is shown in Fig 5 (from Ref 19) The accom-panying analysis indicated that the ideal performance is essentiallythe same as that of a conventional rocket engine as long as the aver-age chamber conditions and fuels used were the same Furthermoreif there was signi cant mixing between the burned and unburnedgases the performance could actually be degraded19 Further workwas done to extend the analysis to two-phase detonations20 and toprovide an explanation of some cases of shock-related combustioninstabilitiesobserved in liquid rocket motors21 However multicycleexperimental operation does not appear to have been achieved

Oblique Shock-Induced DetonationsThe work of Gross11 Gross and Chinitz12 Nicholls et al13

Nicholls and Dabora14 and Dunlap et al15 led to an extensive ex-

Fig 6 Schematic of a normal shock-induced combustion (from Ref 6)

Fig 7 Schematic of an oblique shock-induced combustion (fromRef 6)

perimental study using a Mach 3 tunnel at the Arnold Engineer-ing Development Center A detailed review of this work duringthe period 1959ndash1968 is available6 and hence only a brief discus-sion is given here The shock bottle experiments of Nicholls andDabora14 (Fig 2) were reproduced and a different con gurationwas developed In this con guration [Fig 6 (from Ref 6)] combus-tion takes place behind a normal shock generated by oblique shocksinduced by wedges It has been argued that a detonation does nottake place in this con guration because the normal shock wave isindependently generated by the wedges and is not directly affectedby the combustion6 The term ldquoshock-induced combustionrdquo was in-troduced to describe such phenomena Whether a detonation occursor not their work led to the realization that a normal shock-induceddetonation a ChapmanndashJouguet (CndashJ) detonation was not essentialfor propulsion6 They investigatedother con gurations [Fig 7 (fromRef 6)] and concluded that oblique-shock-induced combustion isexperimentally feasible This was a signi cant conclusion becauseit has led to the future development of the ODWE and to someextent to the ram acceleratorBoth of these propulsion applicationswill be discussed later in some detail

Detonation ThrustersIn the 1970s a substantial effort22 iexcl 25 was undertaken at the Jet

Propulsion Laboratory (JPL) to investigate the feasibility of usingdetonative propulsion for thrusters in the dense or high-pressure at-mospheres of solar system planets The primary reason for studyingthis concept was the potential for high performance for situationswhere the external pressures are high Under such conditions con-ventional thrusters become inef cient due to the lowering of thecombustor to ambient pressure ratio The JPL concept involved ll-ing nozzles with hydrogen helium nitrogen or carbon dioxide tosimulate planetary atmospheres and then intermittentlydetonating asmall quantity of condensed-phase explosives at the apex of the noz-zle Thrust is produced primarily from the momentum of the llergases set into motion by the blast wave and by the expansion ofthe detonation products Both theoretical analysis and experimentsshowed that a small quantity of explosives detonated within the noz-zle lled with gases was suf cient to produce high speci c impulseDifferent types of nozzles and gases of different molecular weightswere utilized With a short-plug nozzle there was a slight reductionin speci c impulse with increasing ambient pressure but the resultswere virtually independent of the molecular weight of the ambi-ent gases However with a long-cone nozzle there was a progres-sive increase in the speci c impulse with increasing ambient pres-sure for high molecular weight gases (carbon dioxide and nitrogen)and a decrease with increasing pressure for lower molecular weightgases (helium) A primary conclusion from the work at JPL was that

KAILASANATH 1701

detonation propulsion technology can be considered for producingboth large velocity changes as a main propulsion system or for gen-erating small velocity changes such as in attitudecontrol propulsionAlthough this can be considered to be one of the more successfulapplications of detonation to propulsion one must remember thatin this application the details of the detonation process itself ap-pear to be secondary and the working uid is a nondetonable gasmixture

Laser-Supported Detonations (LSDs)Another intriguing concept proposed in the 1970s was the use of

lasers to propel objects to low Earth orbit26 The basic idea behindthis propulsion concept is to use lasers to generate detonations anda high kinetic energy propulsive jet moving away from the objectbeing propelled The idea was not pursued very far due to the limitedlaser powers available then and the obvious atmospheric transmis-sion problems However with the planned development of largerlasers under the Strategic Defense Initiative program the conceptwas explored in some detail (for example see Refs 27ndash29) in thelate 1980s In principle a high controllable exhaust velocity can beachieved by a two-step process In the rst step the desired amountof propellant is formed as a gas layer near the surface and in thesecond step the laser energy is directly deposited in this layer The rst layer can be generated either by transpirationof a liquid througha porous surface or by a laser prepulse in a two-pulse system27 Ina two-pulse system a very low-intensity long-duration (on the or-der of seconds) laser prepulse is used to ablate enough gas fromthe surface of the solidliquid propellant to form a buffer gas layerA second high-intensity short-duration laser pulse irradiates thegas layer depositing its energy in the gas layer just before reachingthe propellant surface A shock wave then propagates out throughthe gas layer back toward the laser heating and accelerating thegas layer What is effectively a laser-supported detonation (LSD)is formed by the shock wave followed by the recombination zonebehind it28 In theory a high speci c impulse ( laquo 1000 s) can begenerated with simple propellants such as ice However the realproblems of building a large enough laser and safely transmittingenough energy to the desired location remain and to the best ofthe authorrsquos knowledge LSDs have not been conclusively demon-strated

Another concept using lasers to initiate detonations was proposedby Carrier et al30 Fendell et al31 and Carrier et al32 They con-ducted theoretical and experimental feasibility studies of a super-sonic combustor based on a stabilized conically con gured obliquedetonation wave (ODW) The stabilized detonation wave is formedby the interaction of a train of spherical detonation waves each di-rectly initiated by a brief localized deposition of energy from a veryrapidly repeated pulsed laser The laser is focused on a xed site inthe combustor where there is a steady uniform supersonic streamof detonable gases Initiation of an individual spherical detonationwave by a single laser pulse was achieved32 but further results havenot been reported

ODWEsWith the renewed emphasis on hypersonic ight vehicles in the

1980s extensive investigations were conducted of supersonic com-bustion concepts However these studies will not be discussed hereunless they speci callyinvoke detonations Various versionsof theseconcepts have been called by special names such as detonation-driven ramjets (dramjets)5 but these essentially involve stabilizedODWs and can be traced to the original concept proposed in the1950s by Dunlap et al15 Basically if a premixed fuelndashair mixture ows at a velocity greater than the CndashJ velocity then a normal det-onation wave cannot be stabilized The hope was that under theseconditions an ODW can be stabilized on a wedge-shaped objectsuch that only the normal component of the ow becomes sonic orsubsonic That is most of the ow in the combustor will remainsupersonic A schematic of the dramjet is shown in Fig 8 (fromRef 5) Here fuel is injected into the incoming airstream mixesrapidly with it and is ignited behind the centerbody of a ramjetwhich acts as a stabilizer holding the oblique wave It is similar toa conventional ramjet except for the close coupling between the re-

Fig 8 Schematic of a dramjet (from Ref 5)

Fig 9 Schematic of shock structure for scramjet engine and an ODWE(from Ref 33)

action waves and the leading shock waves resulting in a detonationwave The performance of such a device has been calculated andshown5 to be less than that of an ideal ramjet over a range of Machnumbers from 5 to 30 Issues such as the presence of inlet shocksand the ability to mix rapidly enough to obtain a detonable mixturewere not addressed in this paper5

A representation closer to scramjet engines is shown in Fig 9(from Ref 33) Here the inlet shock system is shown but the linkbetween the inlet shocks and the ODW is left unclear Once againit is assumed that fuelndashair mixing occurs rapidly enough and thatthe oblique shock wave is strong enough so that chemical reactionsoccur close behind it and are coupled to it Analysis of a conceptualvehicle using such an ODWE showed that it has better performancethan an equivalent scramjet for Mach numbers over 15 The mainadvantage arises from the reduced length of detonation engines In arelatedwork the issueof mixing inhomogeneities and their effectonthe detonation wave were addressed numerically34 There have alsobeen severalother studies examining the feasibilityand performanceof detonation wave ramjets3536 All of these concepts essentiallyrely on rapid ignition and combustion occurring behind an obliqueshock wave and the coupling between the two processes resultingin a stable structure A difference between these more recent worksand the earlier ones is the realization that in many situations shock-induced combustion does not lead to detonations However none ofthese studies seem to have led to the actual building and testing ofan ODWE

Ram AcceleratorsIf the solid body anchoring the oblique shock or detonation is

free to move the integral of pressure forces around the body canresult in a net force causing the body to accelerate or decelerateThis is essentially the concept behind the ram accelerator a propul-sion concept proposed3738 in the late 1980s In a ram accelerator

1702 KAILASANATH

Fig 10 Schematic of a ram accelerator in the detonative mode

Fig 11 Early concept of the detonative mode (from Ref 37)

a projectile that resembles the centerbody of a ramjet travels at asupersonic speed through a premixed fuelndashoxidizerndashdiluent mix-ture enclosed in a tube Because the projectile travels at supersonicspeeds oblique shock waves are formed at the nose of the projec-tile and they subsequently re ect from the side walls of the tubeand the projectile body Depending on the strength of these obliqueshocks ignition may occur after the initial shock or subsequentre ections Depending on the velocity of the projectile differentmodes of combustion and correspondingly different modes of op-eration of the system are possible When the velocity of the pro-jectile is greater than the CndashJ velocity an ODW may be stabilizedon the projectile resulting in the detonative mode of operation Aschematic of the ram accelerator operating in the detonative modeis shown in Fig 10 From the early days detonation has been dis-cussed as a means to attain very high velocities including escapevelocities37 In Ref 37 and several subsequent investigations thestructure of the detonative mode was described by the schematicdiagram shown in Fig 11 The projectile nose cone angle and thesound speed in the mixture selected were tailored to avoid combus-tion behind the initial shock wave but to provide initiation behindthe rst shock wave re ected from the tube wall Emphasis was alsoplaced on having the rst re ected shock wave strike the projec-tile body just aft of the shoulder37 If the system needed to be de-signed so precisely successful operation would indeed be dif cultin practical situations Further research has shown that the systemis far more stable and the particular scenario shown in Fig 11 ac-tually occurs just prior to system failure There were early reportsof successful attainment of superdetonative projectile velocities39

but the experimental emphasis shifted to the successful develop-ment of subdetonative modes of operation except at the Instituteof Saint-Louis (ISL) where research on the detonative mode hascontinued40 iexcl 42

Numerical simulations4344 and theoretical studies4546 based onsimpli ed analytical models of superdetonative ram acceleratorswere startedin the earlydays and have made asigni cant impactoverthe past decade These early studies con rmed the potential of ramaccelerators to accelerate ef ciently large masses to escape veloci-ties and low Earth orbit The steady-state simulations of Yungster47

highlighted the role of the projectile velocity (or Mach number)and showed that detonation could occur at different locations alongthe projectile body depending on the projectile Mach number (de- ned as the ratio of the velocity of the projectile to the speed ofsound in the premixed mixture) This work also raised the issue ofcombustion in the boundary layers along the projectile

The projectile velocity changes continuously in the ram acceler-ator and therefore steady-state simulations can at best be thoughtof as trying to characterize an instantaneous snapshot of the sys-tem The importance of the transient processes was rst highlightedin the work of Li et al48 This as well as subsequent work49 hasshown the importance of taking into account the time-dependent

Fig 12 Schematic showing the detailed structure of an oblique deto-nation and its relation to the detonation structures in ram accelerators

nature of the ow in the ram accelerator in describing phenomenasuch as unstarts and other system failures Computational studiesof the starting aspects of ram accelerators and the role of viscouseffects have also been conducted50 iexcl 52 Experimental efforts haveprimarily focused on the nondetonative mode except at ISL40 iexcl 42

Before discussing some of the experimental results and the dif cul-ties encountered in the practical development of ram accelerators inthe detonative mode two critical issues need to be discussed Oneissue is the actual structure of ODWs and the other is the stabilityof such detonation waves during transient operation

Structure and Stability of ODWsA key issue in the development of the ODWE and the ram accel-

erator is the structure and stability of ODWs Although the use ofODW was suggested as early as 195815 the stability of such waveswas implicitly assumed and questions about their detailed structurewere not considered Renewed interest in the late 1980s led to a the-oretical analysis of the supersonic ow of a combustible gas mixturepast a wedge53 Reference 53 indicated that for approach velocitiesroughly 25 or more larger than the CndashJ velocity of the reactantmixture ODWs can be stabilized over a useful range of wedge an-gles For wedge angles greater than a critical value the wave willdetach from the wedge and form an overdriven detonation wave ora more complex shock-induced combustion wave About this timethere were also numerical studies that showed that standing det-onation waves can be established on a wedge under certain owconditions5354

The early experimental efforts61314 were either focused on stabi-lizing a normal detonation wave in a duct or involved shock-inducedcombustion behind weak oblique shock waves Therefore it was notclear if ODWs could really be stabilized until some experiments inthe late 1980s on the diffraction and transmission of detonationsshowed that they could be5556

More recently the interest in ram accelerators has spawned alarge number of studies of the structure and stability of ODWs57 iexcl 61

The work of Li et al57 iexcl 59 showed that the basic structure of anoblique detonation generated by a wedge is more complex thanpreviously thought and consists of a nonreactive shock wave an in-duction region de agration waves and a detonation wave in whichthe pressure increase due to the energy release is closely coupledto the shock front This basic structure and a schematic of how itmight relate to the detonation waves in ram accelerator is shown inFig 12 These studies also showed that over a wide range of ow

KAILASANATH 1703

and mixture conditions the basic detonation structure is stable andvery resilient to disturbances in the ow In agreement with theory53

the simulations also showed the existence of a critical angle beyondwhich an attached stable detonation does not occur59 Other recentstudies have shown the same complex structure even with the inclu-sion of more complex chemical kinetic schemes6061 The essentialfeaturesof the complex ODW structure have also been con rmed byexperiments462 Thus a major conclusion from the recent researchis that the structure of an ODW can no longer be assumed to bea straight oblique shock wave followed closely by energy releaseand the complex structure can indeed be stabilizedon wedge-shapedobjects Implications of this work on propulsion applications mustbe considered

Although the schematic of Fig 12 shows how to relate the basicODW structure to that in ram accelerators the actual detonationwave structure in ram accelerators is likely to be more complexbecause of the expansion waves and multiple shock waves that aregenerated due to the geometric complexity of the system63 Theresults from a study of the detailed structure of detonations in ramaccelerators under a variety of ow conditions63 are summarizedin Fig 13 where the detonation wave structure is shown for three ow Mach numbers Recent experimental investigations have alsorevealed extremely complex detonation wave structures involvingmultiple shock waves under certain conditions64 In spite of theircomplexity all of these structures appear to be inherently relatedto the basic three-wave structure of oblique detonations shown inFig 12 However this appears to be an area of research that willcontinue for a while

a)

b)

c)

Fig 13 Detailed structure of detonations inram accelerators operatingat a) Mach 76 b) Mach 80 and c) Mach 84 (from Ref 63)

a) Case A Mach = 905

b) Case B Mach = 915

Fig 14 Visualization of the ow eld during the acceleration of theprojectile shown in Fig 5 from Mach 8 to 10 at two instants

Transient Nature of Detonation Waves in Ram AcceleratorsThe second important feature revealed by the recent studies is

the inherent transient nature of detonation waves in con gurationssuch as the ram accelerator When the concept of ram acceleratorsin the detonative mode was rst proposed it was thought that theleading shock wave (from the nose of the projectile) should re ectfrom the tube walls such that it impinges exactly at the shoulderof the projectile This will maximize the thrust and hence accel-eration However if the system were designed so precisely for onecondition its off-design performance could be unacceptable Fur-thermore simulations have shown that as the projectile acceleratesthe detonation wave structure tends to also move49 To illustrate thedynamics of detonations in the ram accelerator two photographsof the ow eld from a simulation (where the projectile acceleratedfrom Mach 8 to 10) are shown in Fig 14 In case A the projectileMach number is 905 and in case B it is 915 For both cases themain illustrations show the concentration of water vapor with thetwo inserts showing the pressure and temperature in a selected re-gion of the ow eld At the instant shown in Fig 14a (case A) thepressure and temperature behind the rst re ected shock wave arenot very high and combustion and energy release (as indicated bythe water vapor concentration) takes place only behind the secondre ected shock wave That is the detonation can be said to occurbehind the second re ected shock wave By the time the projectilehas accelerated to Mach 915 the leading shock wave has strength-ened raising the pressure and temperature behind the rst re ectedshock wave and causing the detonation to occur there

The behavior just discussedis seenthroughout the operationof theram accelerator in the detonative mode That is detonation waves inram accelerators are dynamic and move from shock wave to shockwave as the projectile acceleratesEventually the detonation wouldoccur just after the rst shock wave and the whole system will ceaseto operate because there will not be any positive thrust Furthermorethe detonation wave will run in front of the projectile creating a sit-uation similar to the classical unstart phenomena in ramjets Suchunstarts and detonations occurring after a different number of re- ections have also been noted in experiments42 Hence the shape ofthe projectile has to be chosen carefully so that the detonation canmove over the projectile body and still produce positive thrust overthe regime of interest A practical problem that has hindered thedevelopment of ram accelerators operating in the detonative modeis the change in the shape of the projectile due to material erosionburning and failure This is currently being overcome by the use ofsteel cowlings42

1704 KAILASANATH

PDEsThe other major current propulsion application involving detona-

tion waves is the PDE The basic concept behind a PDE is simpleDetonation is initiated repeatedly at either the closed or the openend of a detonation chamber that is lled with a premixed fuelndashairmixture Figure 15 (based on Ref 65) illustrates the concept forthe case when the detonation wave is initiated at the closed endAs shown in this schematic a one-dimensional planar detonation isinitiated near the closed end of the detonation chamber and travelsat the CndashJ velocity toward the open end A set of rarefaction wavestrail the detonation wave to reduce the velocity to zero at the closedend of the tube When the planar detonation wave exits the chamberanother set of expansion waves is generated that travels toward theclosed end These waves evacuate the burned detonation productsresulting in a cool empty chamber that is ready to be lled with afresh fuelndashair mixture The entire cycle repeats resulting in a pe-riodic high-pressure zone near the closed end of the chamber Theintegrated effectof this high pressure over the closed end (represent-ing a thrust wall) produces thrust This is of course a very idealizedpicture of the system Before discussing progress made in the re-cent years in developing a PDE a brief review of the background isprovided

The early work on the use of intermittent detonations has alreadybeen discussed In the late 1980s this concept of using intermittentor pulsed detonations was reexamined experimentally at the NavalPostgraduate School66 An ethylenendashoxygen detonation wave in asmall-diameter tube was used as a predetonator to initiate detona-tions in a larger tube containing an ethylenendashair mixture Periodicfuel injection within the naturally aspirated tube resulted in an in-termittent frequency of 25 Hz Speci c impulse estimated usingthe pressure time history and the amount of fuel consumed rangedfrom 1000 to 1400 s The velocities of the observed waves (lessthan 1 kms) are signi cantly below the CndashJ detonation wave ve-locities for the reported mixtures indicating that a fully developeddetonation wave was not formed

The rst purely computational study of the PDE reported in theopen literature appears to be the work of Cambier and Adelman67

The system simulated consisted of a 50-cm-long main tube attachedto a 43-cm-long diverging nozzle Quasi-one-dimensional simula-tions using the Euler equations for unsteady ow were carried outwith a total variation diminishing scheme and multistep nite ratekinetics A detonation was initiated at the closed end of a tube lledwith a premixed stoichiometric hydrogenndashair mixture at 3 atmOverall performance calculated by integrating the instantaneousthrust and the fuel ow rates gave a speci c impulse Isp of about6500 s and a range of operating frequencies up to 667 Hz Issues ofignition energies and transition to detonation were not consideredand the dif culties with attaining rapid mixing of fuel and air wasavoided by using a premixed fuelndashair mixture These assumptionsare still typical of those used in more recent simulations

The rst look at the interactions between the ow inside the cham-ber and that outside appears to be the work of Eidelman et al7

who simulated a cylindrical detonation chamber (about 15-cm diam

Fig 15 Schematic of an idealized PDE showing various stages duringone cycle of operation (adapted from Ref 65)

Fig 16 Generic PDE device used for somecomputational studieswheredetonation is initiated near the open end (from Ref 69)

and 15 cm long) with a small converging nozzle In their two-dimensional simulations a self-similarsolution for a planar detona-tion was imposed near the open end and made to travel toward theclosed end The re ection of the wave from the closed end createda higher pressure at the wall than during initiation near the closedend Hence this mode of initiation was considered to be better Af-ter the detonation left the chamber the pressure in the chamber wasobserved to fall below the ambient resulting in ingestion of outsideair In subsequent works86869 the same numerical techniques andmodels were used to investigate a slightly different con gurationwhere the nozzle at the aft end was removed and an inlet was in-troduced near the closed end This con guration has become oneof the popular systems for two-dimensional (and axisymmetric) nu-merical simulations Again detonation was assumed to be initiatedat the open end and traveled toward the closed end In this con gu-ration when the pressure falls below atmospheric in the detonationchamber air is ingested through the inlets in the front even duringstatic operations Thus a valveless self-aspiratingoperation is pos-sible These simulations also introduced the discussion of open-endinitiation vs closed-end initiation and their relative advantages Thecurrent status of this issue will be discussed later

Many of the early efforts on intermittent or PDEs have been re-viewed in Ref 7 and 8 where a link is also made between thesedetonation engines and pulse jet engines (which do not use detona-tive combustion) In addition Ref 8 provides a detailed summaryof the numerical investigations conducted by the authorsrsquo group ona generic PDE device shown in Fig 16 A characteristic feature ofthis device is that the detonation wave is initiated at the open endand travels toward a thrust wall at the closed end An advantage ofthis approach is that it is valveless and self-aspirating Air is en-trained into the chamber through inlets near the closed end whenthe pressure falls below atmospheric pressure during part of the cy-cle Unsteady Euler simulations were used to study the operationof the device from static (M = 0) to supersonic (M = 2) conditionsAs stated in Ref 8 a proof of principle experimental demonstra-tion of the PDE mode of operation where detonation is initiated atthe open end still needed to be done Furthermore issues such asdetonation initiation and structure and fuelndashair injection and mixingalso needed to be studied

Bussing and Pappas65 provided a detailed description of thebasic operation of an ideal zed PDE They also reported someone-dimensional studies of PDEs burning hydrogenndashoxygen andhydrogenndashair mixtures The detonation wave was initiated at theclosedend in this engine using a high-temperature and high-pressureregion An advantage of this design is the con nement provided atthe closed end which should enable a more rapid initiation andestablishment of a detonation Both airbreathing and rocket modeof operation of this valved design was discussed Lynch et al70

presented a computational study of an axisymmetric PDE with astraight inlet very similar to that of Eidelman and Grossmann8 andEidelman et al69 Various inlet lengths from 1 to 12 cm and ightMach numbers of 08 and 20 were studied The results were similarto those reported earlier except that some additional details of thedetonation wave front and the air induction system were resolved inthese studies In a related work Lynch and Edelman71 showed thatreductions in scavenging time (and hence overall cycle time) canbe achieved by suitably shaping the inlet

Bussing et al72 compared open-end initiation and closed-end ini-tiation in a simple tube PDE geometry Equivalent thrust productionand fuel ef ciency were observed for the two types of initiation

KAILASANATH 1705

Although a higher head-end pressure was attained during open-endinitiation the pressure continuously dropped for that case unlike theclosed-end initiation where a constant pressure region occurred fora while before reduction of the pressure to the ambient value Morerecently an experimental study has con rmed the differences in theobserved pressure pro les and it has also been concluded that themaximum level of impulse attained is independent of the locationof the detonation initiator73

In the past few years there has been a substantial growth in theappearance of papers related to the PDE and complete sessionshave been devoted to the topic at annual AIAAASMESAEASEEjoint propulsion conferences Because of space considerations onlyselected papers from these years are brie y discussed here Sterlinget al74 conducted one-dimensional numerical investigations of aself-aspiratingPDE operating for multiple cyclesA key observationfrom their studies was that only a portion of the detonation tubecan be lled with fresh charge under self-aspirating operation Inspite of this limitation they reported a speci c impulse of 5151 sfor a hydrogenndashair system However they concluded that the idealperformance of such an engine is ldquonear those of other hydrogen-fueledair breathing enginesrdquo74

Several basic shock tube experiments related to the developmentof a hydrogen-fueled PDE were described by Hinkey et al75 Theseinclude the measurement of detonation wave velocities and de a-gration to detonation transition (DDT) lengths for a range of equiv-alence ratios As expected the DDT lengths were found to be toolarge for practical applications even in hydrogenndashoxygen mixturesTherefore traditional techniques such as the use of a Schelkin spiral(see Ref 76) were tried to reduce the transition length A factor of2ndash4 reduction was observed over the range of equivalence ratios in-vestigated This work75 highlights the dif culties involved in obtain-ing a fully developed detonation in a short distance (as assumed inthe conceptual studies and numerical simulations discussed earlier)and questions if the detonations reported in previous experimentalinvestigations161766 were really fully transitioned from de agra-tions The speci c impulses measured in their experiments (about240 s for hydrogenndashoxygen and 1200 s for hydrogenndashair) were alsostated to be in agreement with their analysis

An interesting PDE concept presented by Bussing77 was that of arotary-valved multiple-pulsed detonation engine Here several det-onation chambers are coupled to an air inlet and fuel source usinga rotary valve as suggested in one of the early papers of Nichollset al16 The rotary valves allow the lling of some detonation cham-bers while others are detonating or exhausting Other papers includepresentation of an overall PDE performance model78 and the exper-imental characterization of the detonation properties of some fuelsfor use in PDEs79

The importance of adequately mixing the fuel and oxidizer washighlighted by the experimental investigations of Stanley et al80

who obtained very low sub-CndashJ velocities when injecting the fueland oxidizer at different times and not taking extra effort to en-sure that they were well mixed The use of turbulence producingdevices appeared to improve signi cantly the mixing and the at-tainment of higher velocities but also resulted in signi cant thrustlosses This study also showed that raising the initial pressure wasbene cial The computational studies of Eidelman et al81 showedthat the PDE engine could operate even for a range of transitionaldetonation regimes that produced nonplanar or not fully developeddetonations Aarnio et al82 discussed two failure modes DDT tran-sition failure and premature ignition In both cases the system con-tinued to operate though there would be loss of thrust In Ref 82detailed discussions of the pressure and thrust histories during a5-Hz 20-cycle operation are also provided (Fig 17) The cycle tocycle variations of peak pressure have been attributed to the lowsampling frequency used82 Both a direct measurement as well asthe integration of the pressure histories were used to estimate the av-erage Isp from hydrogenndashair systems to be between 1116 and 1333 sThe effect of the predetonator volume on the reduction of speci cimpulse was also shown to be a critical issue The effects of inletsand nozzles were not considered in this study

A conceptual design of a multitube PDE with a single air inletduct was discussed by Pegg et al83 Time-dependent computational

Fig 17 Pressure history at one location from a PDE operating at 5 Hzfor 20 cycles (from Ref 82)

uid dynamics analysis indicated that the inlet isolaterdiffusercon-cept would work and did not allow enough time for the formationof destabilizing hammer shocks Performance analysis includingcomponent ef ciencies reported in the paper indicate that opera-tional frequencies of the order of 75ndash100 Hz are required for theMach 12ndash3 ights considered Experimental data indicating mul-ticycle operation at 100 Hz with a rich ethylenendashoxygen mixturewere presented by Sterling et al84 DDT enhancement devices andpredetonators were used in the experiments In spite of this det-onation waves did not occur all of the time and when detonationwaves failed the tube became very hot and testing was suspendedTherefore the maximum testing time was reported to be 05 s Inspite of this limitation the two studies8384 taken together high-light the possibility of PDEs burning hydrocarbon fuels for realisticmissions

Five different nozzle shapes and their effect on performance werestudied computationally by Cambier and Tegner85 Their results in-dicate that the presence of a nozzle can affect the performance ofthe PDE by increasing the thrust delivery during the ignition phaseThe bell-shaped nozzles appeared to give higher performance thanshapes with positive curvature Their results also showed that noz-zles also affected the ow dynamics and hence the timing of thevarious phases of the engine cycle In an experimental study of aconical exhaust nozzle the authors of Ref 86 noted improved per-formance (based on higher velocities which were still sub-CndashJ) butdid not see any effect on the blowdown process More recently87

the effectof various nozzle shapes including converging divergingand straight have been reexamined computationally The converg-ing sections of nozzles introduced shock wave re ections whereasdiverging sections generated a negative thrust for a portion of thecycle due to overexpansion In spite of these limitations the overallconclusion of these studies was that ldquonozzles can drastically in-crease ef ciency of the PDEsrdquo87 There is also some experimentalevidence suggesting that propulsive performances can be increasedby the addition of a nozzle73 However factors such as the effect ofthe nozzles on detonation transmission and the detailed dynamicsof the ow have not yet been elucidated This appears to be an areathat needs further investigation

Other recent efforts have focused on demonstrating the opera-tion of single- and multitube combustors coupled to an inlet using arotary-valvemechanism88 The valve servesto both meter the air owinto the combustor and to isolate the inlet from the high pressuresproduced during the detonation cycle Utilizing multiple combus-tors that ll and detonate out of phase allows the continual use of thein owing air Firing rates of up to 12 Hz per combustor were demon-strated for a hydrogen-fueled system Several other issues related tothe practical development of PDEs have been discussed by Bussinget al89 According to them three of the more important engineer-ing issues are the development of a valve-injection subsystem anignition system and a low-volume predetonator

The use of PDEs for rockets has been revived recently90 Therocket mode of operation is very similar to the description providedearlier of the airbreathing engine with ignition at the closed endexcept that the oxidizer also needs to be injected into the system pe-riodically An advantage of PDEs for rockets would be their higherpower density thus enabling the development of more compact

1706 KAILASANATH

rockets Reference 90 shows the pressure traces from a pulsed det-onation rocket engine operating at 145 Hz on a hydrogenndashoxygenmixture

More recently a two-tube rotary-valved PDE with ight-sizecomponents was operated at 40 Hz per combustor for 30 s withan ethylenendashair mixture91 In addition throttling of the device al-lowed operation at three thrust levels during a 10-s demonstrationexperiment Contrary to some earlier estimates the thermal loadswere signi cant and it was not possible to operate for more than3 s without water cooling It is not clear if this was because ofperiodic detonation failures and consequent de agrative modes ofcombustion as reported in some earlier studies

The work reported in the open literature and discussed earlier hasfocused on gaseous fuels However for volume-limited propulsionapplications PDEs operating on liquid fuels need to be demon-strated In a recent study Brophy et al92 reported on experimentsusing JP-10oxygen and JP-10air aerosols The fuelndashoxygen mix-ture was successfully detonated to obtain an engine operating at5 Hz but the tests with the fuelndashair mixture were not successful

Computational studies continue to make progress in the study ofthe PDE8793 iexcl 97 The effect of nozzle shapes and incomplete transi-tion to CndashJ detonations have been studied by Eidelman and Yang87

They nd that the cycle ef ciency will be virtually the same whethera CndashJ detonation wave is initiated instantaneously or if the CndashJ valueis attainedonly at the end of the detonation chamber Primarilybasedon this study87 they come to an interesting conclusion that ldquoinitia-tion energies that are required for PDE operations can be small andcomparable to energies required for initiation of (other) combustionprocessesrdquo However this is yet to be substantiated experimentallySekar et al93 reexamine open- and closed-end initiation and thebasic self-aspirating con guration of Eidelman8 using more recentcomputational tools Injection and mixing issues that have gener-ally been ignored in previous numerical studies are also beginningto be addressed94 Both front-wall and lateral-wall injections werestudied and the time required for fuel injection was estimated Itis suggested that it may be possible to ll only a portion of thechamber during high-frequency operations Recent computationalstudies also suggest that preconditioning the fuelndashair mixture us-ing a shock wave can signi cantly reduce the DDT distance95 Theperformance estimates mentioned throughout the discussion of thePDE show signi cant variation A systematic review of various per-formance estimates and possible reasons for the observed variationhave been presented9697 These numerical studies suggest that thetime history of the back pressure is a crucial parameter affecting theestimated performance

Other Detonation EnginesThere have also been several other interesting conceptual studies

on the applications of detonations to engines Cambier et al98 haveproposed a pulsed-detonation wave augmentation device as part ofa hybrid engine for a single-stage-to-orbit airbreathing hypersonicvehicle Here the PDE is used for both thrust generation as well asfor mixingcombustion augmentation Two speci c con gurationswere investigated numerically

The concept of a detonation internal combustion engine was pro-posed by Loth and Loth99 and Loth100 Here a conventional pistoncylinder con guration is modi ed to incorporate a separate deto-nation chamber which is isolated by a valve from the compressionchamber during fuel injection and discharges tangentially into a vor-texchamber formedby the piston and cylinder at top deadcenterTherapid detonative combustion is followed by rapid dilution throughmixing with air in the vortex chamber to reduce the formation ofNOx and unburned hydrocarbons and to achieve an overall leanmixture The vortex chamber was also designed to store a portion ofthe detonation waversquos kinetic energy Overall thermal ef cienciessomewhat better than constant volume combustion were postulatedunder ideal conditions

ConclusionsDetonation waves have been explored extensively for propulsion

applications because of their inherent theoretical advantage overde agrative combustion However practical developments to date

have been of idealized systems or laboratory-scale devices or bothThe basic ideas behind most of the current systems being investi-gated have been known for a while but still many of the details needto be understood Advances in computations and experimental di-agnostics appear to be poised to make practical propulsion devicesbased on detonation waves a reality

Two major systems currently under development are the ram ac-celerator and the PDE In both cases the theoretical and computa-tional studies have been far more encouraging than the experimentsHowever in both cases actual demonstrations of devices workingon gaseous fuels have taken place In the case of the ram accelera-tor fuelndashoxygenndashdiluent mixtures have been used and the primaryproblem appears to be heat transfer and related material failuresThe use of gaseous fuels is not a major issue here but new materialsfor the projectiles need to be explored

In the case of the PDE most of the successful operations havebeen with fuelndashoxygen mixtures There still has not been a directcomparison between detailed experimental data and related theoret-ical and computational results Furthermore most of the PDE andrelated work to date have focused on gaseous mixtures PDEs op-erating on multiphase mixtures need to be emphasized because itwould be more practical to use liquid fuels for most of the propul-sion applicationsproposed for these systemsThere are several othergeneral issues that need to be resolved Clearly the evacuation ofexhaust gases is important both from the point of view of controllingthe cycle time and for preventing autoignition and mixing with freshgases The heat transfer and noise from PDEs have also not beenreported in any detail Other issues include inletndashcombustor andcombustorndashnozzle interactions and scalability of the entire device

AcknowledgmentsThis work has been sponsored by the Of ce of Naval Research

through the Mechanics and Energy Conversion Division and theUS Naval Research Laboratory

References1Fickett W and Davis W C Detonation Univ of California Press

Berkeley CA 19792Hoffmann N ldquoReaction Propulsion by Intermittent Detonative Com-

bustionrdquo German Ministry of Supply AI152365 Volkenrode Translation1940 (cited in Ref 16)

3Roy M ldquoPropulsion par Statoreacteur a Detonationrdquo Comptes-Rendusde lrsquoAcademie des Sciences Vol 222 1946 pp 31 32

4Dabora E K and Broda J C ldquoStanding Normal Detonations andOblique Detonations for Propulsionrdquo AIAA Paper 93-2325 July 1993

5Dabora E K ldquoStatus of Gaseous Detonation Waves and Their Rolein Propulsionrdquo Fall Technical Meeting of the Eastern States Section ofthe Combustion Institute Combustion Inst Pittsburgh PA 1994 pp 11ndash

186Rubins P M and Bauer R C ldquoReview of Shock-Induced Supersonic

Combustion Research and Hypersonic Applicationsrdquo Journal of Propulsionand Power Vol 10 No 5 1994 pp 593ndash601

7Eidelman S Grossmann W and Lottati I ldquoReview of PropulsionApplications and Numerical Simulations of the Pulsed Detonation EngineConceptrdquo Journal of Propulsion and Power Vol 7 No 6 1991 pp 857ndash

865 also AIAA Paper 89-2446 July 19898Eidelman S and Grossmann W ldquoPulsed Detonation Engine Experi-

mental and Theoretical Reviewrdquo AIAA Paper 92-3168 July 19929Reingold L ldquoRecherches sur les combustions permanentes apportees

aux foyers a circulation interne supersoniquerdquo ONERA TN2 Dept of En-ergy and Propulsion Study 728-E Paris March 1950

10Bitondo D and Bollay W ldquoPreliminary Performance Analysis ofthe Pulse-Detonation-Jet Engine Systemrdquo Aerophysics Development CorpRept ADC-102-1 April 1952

11Gross R A ldquoExploratory Study of Combustion in Supersonic FlowrdquoUS Air Force Of ce of Scienti c Research TN 59-587 Washington DCJune 1959

12Gross R A and Chinitz W ldquoA Study of Supersonic CombustionrdquoJournal of Aerospace Science Vol 27 No 7 1960 pp 517ndash525

13Nicholls J A Dabora E K and Gealer R L ldquoStudies in Con-nection with Stabilized Gaseous Detonation Wavesrdquo Seventh Symposium(International) on Combustion Combustion Inst Pittsburgh PA 1959 pp766ndash772

14Nicholls J A and Dabora E K ldquoRecent Results on Standing Detona-tion Wavesrdquo Eighth Symposium (International) on Combustion CombustionInst Pittsburgh PA 1962 pp 644ndash655

KAILASANATH 1707

15Dunlap R Brehm R L and Nicholls J A ldquoA Preliminary Studyof the Application of Steady-State Detonative Combustion to a ReactionEnginerdquo Jet Propulsion Vol 28 No 7 1958 pp 451ndash456

16Nicholls J A Wilkinson H R and Morrison R B ldquoIntermittentDetonation as a Thrust-Producing Mechanismrdquo Jet Propulsion Vol 27No 5 1957 pp 534ndash541

17Krzycki L J ldquoPerformance Characteristics of an Intermittent Detona-tion Devicerdquo US Naval Ordnance Test Station US Naval Weapons Rept7655 China Lake CA 1962

18Nicholls J A Cullen R E and Ragland K W ldquoFeasibility Studiesof a Rotating Detonation Wave Rocket Motorrdquo Journal of Spacecraft andRockets Vol 3 No 6 1966 pp 893ndash898

19Adamson T C and Olsson G R ldquoPerformance Analysis of a RotatingDetonation Wave Rocket Enginerdquo Astronautica Acta Vol 13 No 4 1967pp 405ndash415

20Shen P I and Adamson T C ldquoTheoretical Analysis of a RotatingTwo-Phase Detonation in Liquid Rocket Motorsrdquo Astronautica Acta Vol17 No 4 1972 pp 715ndash728

21Clayton R M and Rogero R S ldquoExperimental Measurements on aRotating Detonation-Like Wave Observed During Liquid Rocket ResonantCombustionrdquo Jet Propulsion Lab TR 32-788 California Inst of Technol-ogy Pasadena CA Aug 1965

22Back L H ldquoApplication of Blast Wave Theory to Explosive Propul-sionrdquo Acta Astronautica Vol 2 1975 pp 391ndash407

23Varsi G Back L H and Kim K ldquoBlast Wave in a Nozzle for Propul-sion Applicationsrdquo Acta Astronautica Vol 3 1976 pp 141ndash156

24Kim K Varsi G and Back LH ldquoBlast Wave Analysis for DetonationPropulsionrdquo AIAA Journal Vol 10 No 10 1977 pp 1500ndash1502

25Back L H Dowler W L and Varsi G ldquoDetonation PropulsionExperiments and Theoryrdquo AIAA Journal Vol 21 No 10 1983 pp 1418ndash

142726Kantrowitz A ldquoPropulsion to Orbit by Ground Based Lasersrdquo Astro-

nautics and Aeronautics Vol 10 No 5 1972 p 7427Hyde R A ldquoOne-D Modelling of a Two-Pulse LSD Thrusterrdquo Pro-

ceedings of the 1986 SDIODARPA Workshop on Laser Propulsion editedby J T Kare Lawrence Livermore National Labs 1987 pp 79ndash88

28Emery M H Kailasanath K Oran E S and Gardner J H ldquoCompu-tational Studies of Laser Supported Detonationsrdquo Proceedings of the 1987SDIO Workshop on Laser Propulsion edited by J T Kare Lawrence Liv-ermore National Labs 1990 pp 43ndash55

29Emery M H Kailasanath K and Oran E S ldquoNumerical Simu-lations of Laser Supported Detonationsrdquo 1988 SDIO Workshop on LaserPropulsion Lawrence Livermore National Labs April 1988

30Carrier G F Fendell F McGregor D Cook S and Vazirani MldquoLaser-Initiated Conical Detonation Wave for Supersonic CombustionrdquoAIAA Paper 91-0578 Jan 1991

31Fendell F E Mitchell J McGregor R Magiawala K and Shef eldM ldquoLaser-Initiated Conical Detonation Wave for Supersonic CombustionIIrdquo AIAA Paper 92-0088 Jan 1992

32Carrier G F Fendell F E and Chou M S ldquoLaser-Initiated ConicalDetonation Wave for Supersonic Combustion IIIrdquo AIAA Paper 92-3247July 1992

33Menees G P Adelman H G Cambier J-L and Bowles J V ldquoWaveCombustors for Trans-Atmospheric Vehiclesrdquo Journal of Propulsion andPower Vol 8 No 3 1992 pp 709ndash713

34Cambier J L Adelman H and Menees G ldquoNumerical Simulationsof an Oblique Detonation Wave Enginerdquo AIAA Paper 88-0063 1988

35Atamanchuk T and Sislian J ldquoOn- and Off-design Performance Anal-ysis of Hypersonic Detonation Wave Ramjetsrdquo AIAA Paper 90-2473 1990

36Kuznetsov M M Neyland V J and Sayapin G N ldquoEf ciency In-vestigation of Scramjet with Detonation and Shockless Supersonic Combus-tionrdquo Utchenye Zapisky TsAGI Vol 23 No 2 1992 pp 30ndash37

37Bruckner A P and Hertzberg A ldquoRam Accelerator Direct LaunchSystem for Space Cargordquo International Astronautical Federation Paper 87-211 1987

38Hertzberg A Bruckner A P and Bogdanoff D W ldquoRam Accel-erator A New Chemical Method for Accelerating Projectiles to UltrahighVelocitiesrdquo AIAA Journal Vol 26 No 2 1988 pp 195ndash203

39Kull A Burnham E Knowlen C Bruckner A P and HertzbergA ldquoExperimental Studies of the Superdetonative Ram Accelerator ModesrdquoAIAA Paper 89-2632 1989

40Patz G Seiler F Smeets G and Srulijes J ldquoStatus of ISLrsquos RA-MAC30 with Fin Guided Projectiles Accelerated in a Smooth Borerdquo Pro-ceedings of the Second International Workshop on Ram Accelerators Univof Washington Seattle WA 1995

41Seiler F Patz G Smeets G and Srulijes J ldquoGasdynamic Limitsof Ignition and Combustion of a Gas Mixture in ISLrsquos RAMAC30 ScramAcceleratorrdquo Proceedings of the Second International Workshop on RamAccelerators Univ of Washington Seattle WA 1995

42Seiler F Patz G Smeets G and Srulijes J ldquoIn uence of ProjectileMaterial and Gas Composition on Superdetonative Combustion in ISLrsquosRAMAC30rdquo AIAA Paper 98-3445 July 1998

43Brackett D C and Bogdanoff D W ldquoComputational Investigation ofOblique Detonation Ramjet-in-Tube Conceptsrdquo Journal of Propulsion andPower Vol 5 No 3 1989 pp 276ndash281

44Yungster S Eberhardt S and Bruckner A P ldquoNumerical Simula-tion of Shock-Induced Combustion Generated by High-Speed Projectiles inDetonable Gas Mixturesrdquo AIAA Paper 89-0673 1989

45Rom J and Kivity Y ldquoAccelerating Projectiles Up to 12 kms Utiliz-ing the Continuous Detonation Propulsion Methodrdquo AIAA Paper 88-29691988

46Humphrey J W ldquoParametric Study of an ODW Scramaccelerator forHypersonic Test Facilitiesrdquo AIAA Paper 90-2470 1990

47Yungster S ldquoNavierndashStokes Simulation of the Supersonic CombustionFlow eld in a Ram Acceleratorrdquo AIAA Paper 91-1916 1991

48Li C Kailasanath K Oran E S Boris J P Paper and LandsbergA M ldquoNumerical Simulations of Transient Flows in Ram AcceleratorsrdquoAIAA Paper 93-1916 1993

49Li C Kailasanath K Oran E S Landsberg A M and Boris J PldquoDynamics of Oblique Detonations in Ram Acceleratorsrdquo Shock Waves Vol5 June 1995 pp 97ndash101

50Yungster S and Radhakrishnan K ldquoComputational Study of FlowEstablishment in a Ram Acceleratorrdquo AIAA Paper 98-2489 July 1995

51Choi J Y Jeung I-S and Yoon Y ldquoNumerical Study of ScramAccelerator Starting Characteristicsrdquo AIAA Journal Vol 36 No 6 1998pp 1029ndash1038

52Bezgin L Ganzhelo A Gouskov O and Kopchenov V ldquoSome Nu-merical Investigation Results of Shock-Induced Combustionrdquo Proceedingsof the AIAA 8th International Space Planes and Hypersonic Systems andTechnologies Conference AIAA Reston VA 1998 pp 72ndash82

53Pratt D W Humphrey J W and Glenn D E ldquoMorphology of Stand-ing Oblique Detonation Wavesrdquo Journal of Propulsion and Power Vol 7No 5 1991

54Fujiwara T Matsuo A and Nomoto H ldquoTwo-Dimensional Detona-tion Supported by a Blunt Body or a Wedgerdquo AIAA Paper 88-0089 Jan1988

55Fan B C Sichel M and Kauffman C W ldquoAnalysis of ObliqueShock-Detonation Wave Interactions in the Supersonic Flow of a Com-bustible Mediumrdquo AIAA Paper 88-0441 Jan 1988

56Dabora E K Desbordes D Guerraud C and Wagner H GldquoOblique Detonations at Hypersonic Velocitiesrdquo Dynamics of Detonationsand Explosions Detonations Vol 133 Progress in Astronautics and Aero-nautics AIAA Washington DC 1990 pp 187ndash201

57Li C Kailasanath K and Oran E S ldquoStability of Oblique Detona-tions in Ram Acceleratorsrdquo AIAA Paper 92-0089 Jan 1992

58Li C Kailasanath K and Oran E S ldquoStructure of Reaction WavesBehind Oblique Shocksrdquo Dynamic Aspects of Detonation Vol 153 Progressin Astronautics and Aeronautics AIAA Washington DC 1993 pp 231ndash

24059Li C Kailasanath K and Oran E S ldquoDetonation Structures Behind

Oblique Shocksrdquo Physics of Fluids Vol 6 April 1994 pp 1600ndash161160Da Silva L F and Deshaies B ldquoNumerical Study of Some Ignition

Regimes of Combustible Supersonic Flows over a Wedgerdquo Proceedings ofSecond International Workshop on Ram Accelerators Paper 15 Univ ofWashington Seattle WA 1995

61Thaker A A and Chelliah H K ldquoNumerical Prediction of ObliqueDetonation Wave Structures Using Detailed and Reduced Reaction Mecha-nismsrdquo Combustion Theory and Modelling Vol 1 No 4 1997 pp 347ndash376

62Desbordes D Guerraud C and Hamada L ldquoSupersonic H2-AirCombustions Behind Oblique Shock Wavesrdquo 14th International Colloquiumon Dynamics of Explosions and Reactive Systems Paper D28 Univ ofCoimbra Portugal Aug 1993

63Li C Kailasanath K and Oran E S ldquoDetonation Structures Gener-ated by Multiple Shocks on Ram-Accelerator Projectilesrdquo Combustion andFlame Vol 108 1997 pp 173ndash186

64Viguier C and Desbordes D ldquoConditions of Onset and Stabilizationof Oblique Detonation Wavesrdquo Advances in Experimentation and Compu-tation of Detonations edited by G D Roy S H Frolov K Kailasanathand N Smirnov ENAS Moscow 1998 pp 40 41

65Bussing Tand Pappas G ldquoIntroduction toPulseDetonation EnginesrdquoAIAA Paper 94-0263 1994

66Helman D Shreeve R P and Eidelman S ldquoDetonation Pulse En-ginerdquo AIAA Paper 86-1683 June 1986

67Cambier J L and Adelman H G ldquoPreliminary Numerical Simu-lations of a Pulsed Detonation Wave Enginerdquo AIAA Paper 88-2960 July1988

68Eidelman S Grossmann W and Lottati I ldquoAir-Breathing PulsedDetonationEngine Concept A Numerical Studyrdquo AIAA Paper 90-2420 July1990

69Eidelman S Grossmann W and Lottati I ldquoComputational Analysisof Pulsed Detonation Engines and Applicationsrdquo AIAA Paper 90-0460 Jan1990

1700 KAILASANATH

Fig 4 Comparison of theoretical and experimental results on timehistories of detonation tube acceleration for a 50 acetylenendashoxygendetonation (from Ref 16)

Fig 5 Possible con guration for a rotating detonation wave engine(from Ref 19)

or merely as a pulse jet engine due to the low initiation energies em-ployed His overall conclusion was that although thrust was possiblefrom such a device practical applications were not promising

Rotating DetonationAnother interesting concept that was explored was that of a rotat-

ing detonation wave rocket motor18 iexcl 20 Here gaseous fuel (hydrogenor methane) and oxygen were continuously injected into an annularcombustion chamber and ignited using spark plugs A unidirec-tional detonation wave was created after transition from a de agra-tion wave and the exhaust gases were expelled through an annularnozzle Such a system is shown in Fig 5 (from Ref 19) The accom-panying analysis indicated that the ideal performance is essentiallythe same as that of a conventional rocket engine as long as the aver-age chamber conditions and fuels used were the same Furthermoreif there was signi cant mixing between the burned and unburnedgases the performance could actually be degraded19 Further workwas done to extend the analysis to two-phase detonations20 and toprovide an explanation of some cases of shock-related combustioninstabilitiesobserved in liquid rocket motors21 However multicycleexperimental operation does not appear to have been achieved

Oblique Shock-Induced DetonationsThe work of Gross11 Gross and Chinitz12 Nicholls et al13

Nicholls and Dabora14 and Dunlap et al15 led to an extensive ex-

Fig 6 Schematic of a normal shock-induced combustion (from Ref 6)

Fig 7 Schematic of an oblique shock-induced combustion (fromRef 6)

perimental study using a Mach 3 tunnel at the Arnold Engineer-ing Development Center A detailed review of this work duringthe period 1959ndash1968 is available6 and hence only a brief discus-sion is given here The shock bottle experiments of Nicholls andDabora14 (Fig 2) were reproduced and a different con gurationwas developed In this con guration [Fig 6 (from Ref 6)] combus-tion takes place behind a normal shock generated by oblique shocksinduced by wedges It has been argued that a detonation does nottake place in this con guration because the normal shock wave isindependently generated by the wedges and is not directly affectedby the combustion6 The term ldquoshock-induced combustionrdquo was in-troduced to describe such phenomena Whether a detonation occursor not their work led to the realization that a normal shock-induceddetonation a ChapmanndashJouguet (CndashJ) detonation was not essentialfor propulsion6 They investigatedother con gurations [Fig 7 (fromRef 6)] and concluded that oblique-shock-induced combustion isexperimentally feasible This was a signi cant conclusion becauseit has led to the future development of the ODWE and to someextent to the ram acceleratorBoth of these propulsion applicationswill be discussed later in some detail

Detonation ThrustersIn the 1970s a substantial effort22 iexcl 25 was undertaken at the Jet

Propulsion Laboratory (JPL) to investigate the feasibility of usingdetonative propulsion for thrusters in the dense or high-pressure at-mospheres of solar system planets The primary reason for studyingthis concept was the potential for high performance for situationswhere the external pressures are high Under such conditions con-ventional thrusters become inef cient due to the lowering of thecombustor to ambient pressure ratio The JPL concept involved ll-ing nozzles with hydrogen helium nitrogen or carbon dioxide tosimulate planetary atmospheres and then intermittentlydetonating asmall quantity of condensed-phase explosives at the apex of the noz-zle Thrust is produced primarily from the momentum of the llergases set into motion by the blast wave and by the expansion ofthe detonation products Both theoretical analysis and experimentsshowed that a small quantity of explosives detonated within the noz-zle lled with gases was suf cient to produce high speci c impulseDifferent types of nozzles and gases of different molecular weightswere utilized With a short-plug nozzle there was a slight reductionin speci c impulse with increasing ambient pressure but the resultswere virtually independent of the molecular weight of the ambi-ent gases However with a long-cone nozzle there was a progres-sive increase in the speci c impulse with increasing ambient pres-sure for high molecular weight gases (carbon dioxide and nitrogen)and a decrease with increasing pressure for lower molecular weightgases (helium) A primary conclusion from the work at JPL was that

KAILASANATH 1701

detonation propulsion technology can be considered for producingboth large velocity changes as a main propulsion system or for gen-erating small velocity changes such as in attitudecontrol propulsionAlthough this can be considered to be one of the more successfulapplications of detonation to propulsion one must remember thatin this application the details of the detonation process itself ap-pear to be secondary and the working uid is a nondetonable gasmixture

Laser-Supported Detonations (LSDs)Another intriguing concept proposed in the 1970s was the use of

lasers to propel objects to low Earth orbit26 The basic idea behindthis propulsion concept is to use lasers to generate detonations anda high kinetic energy propulsive jet moving away from the objectbeing propelled The idea was not pursued very far due to the limitedlaser powers available then and the obvious atmospheric transmis-sion problems However with the planned development of largerlasers under the Strategic Defense Initiative program the conceptwas explored in some detail (for example see Refs 27ndash29) in thelate 1980s In principle a high controllable exhaust velocity can beachieved by a two-step process In the rst step the desired amountof propellant is formed as a gas layer near the surface and in thesecond step the laser energy is directly deposited in this layer The rst layer can be generated either by transpirationof a liquid througha porous surface or by a laser prepulse in a two-pulse system27 Ina two-pulse system a very low-intensity long-duration (on the or-der of seconds) laser prepulse is used to ablate enough gas fromthe surface of the solidliquid propellant to form a buffer gas layerA second high-intensity short-duration laser pulse irradiates thegas layer depositing its energy in the gas layer just before reachingthe propellant surface A shock wave then propagates out throughthe gas layer back toward the laser heating and accelerating thegas layer What is effectively a laser-supported detonation (LSD)is formed by the shock wave followed by the recombination zonebehind it28 In theory a high speci c impulse ( laquo 1000 s) can begenerated with simple propellants such as ice However the realproblems of building a large enough laser and safely transmittingenough energy to the desired location remain and to the best ofthe authorrsquos knowledge LSDs have not been conclusively demon-strated

Another concept using lasers to initiate detonations was proposedby Carrier et al30 Fendell et al31 and Carrier et al32 They con-ducted theoretical and experimental feasibility studies of a super-sonic combustor based on a stabilized conically con gured obliquedetonation wave (ODW) The stabilized detonation wave is formedby the interaction of a train of spherical detonation waves each di-rectly initiated by a brief localized deposition of energy from a veryrapidly repeated pulsed laser The laser is focused on a xed site inthe combustor where there is a steady uniform supersonic streamof detonable gases Initiation of an individual spherical detonationwave by a single laser pulse was achieved32 but further results havenot been reported

ODWEsWith the renewed emphasis on hypersonic ight vehicles in the

1980s extensive investigations were conducted of supersonic com-bustion concepts However these studies will not be discussed hereunless they speci callyinvoke detonations Various versionsof theseconcepts have been called by special names such as detonation-driven ramjets (dramjets)5 but these essentially involve stabilizedODWs and can be traced to the original concept proposed in the1950s by Dunlap et al15 Basically if a premixed fuelndashair mixture ows at a velocity greater than the CndashJ velocity then a normal det-onation wave cannot be stabilized The hope was that under theseconditions an ODW can be stabilized on a wedge-shaped objectsuch that only the normal component of the ow becomes sonic orsubsonic That is most of the ow in the combustor will remainsupersonic A schematic of the dramjet is shown in Fig 8 (fromRef 5) Here fuel is injected into the incoming airstream mixesrapidly with it and is ignited behind the centerbody of a ramjetwhich acts as a stabilizer holding the oblique wave It is similar toa conventional ramjet except for the close coupling between the re-

Fig 8 Schematic of a dramjet (from Ref 5)

Fig 9 Schematic of shock structure for scramjet engine and an ODWE(from Ref 33)

action waves and the leading shock waves resulting in a detonationwave The performance of such a device has been calculated andshown5 to be less than that of an ideal ramjet over a range of Machnumbers from 5 to 30 Issues such as the presence of inlet shocksand the ability to mix rapidly enough to obtain a detonable mixturewere not addressed in this paper5

A representation closer to scramjet engines is shown in Fig 9(from Ref 33) Here the inlet shock system is shown but the linkbetween the inlet shocks and the ODW is left unclear Once againit is assumed that fuelndashair mixing occurs rapidly enough and thatthe oblique shock wave is strong enough so that chemical reactionsoccur close behind it and are coupled to it Analysis of a conceptualvehicle using such an ODWE showed that it has better performancethan an equivalent scramjet for Mach numbers over 15 The mainadvantage arises from the reduced length of detonation engines In arelatedwork the issueof mixing inhomogeneities and their effectonthe detonation wave were addressed numerically34 There have alsobeen severalother studies examining the feasibilityand performanceof detonation wave ramjets3536 All of these concepts essentiallyrely on rapid ignition and combustion occurring behind an obliqueshock wave and the coupling between the two processes resultingin a stable structure A difference between these more recent worksand the earlier ones is the realization that in many situations shock-induced combustion does not lead to detonations However none ofthese studies seem to have led to the actual building and testing ofan ODWE

Ram AcceleratorsIf the solid body anchoring the oblique shock or detonation is

free to move the integral of pressure forces around the body canresult in a net force causing the body to accelerate or decelerateThis is essentially the concept behind the ram accelerator a propul-sion concept proposed3738 in the late 1980s In a ram accelerator

1702 KAILASANATH

Fig 10 Schematic of a ram accelerator in the detonative mode

Fig 11 Early concept of the detonative mode (from Ref 37)

a projectile that resembles the centerbody of a ramjet travels at asupersonic speed through a premixed fuelndashoxidizerndashdiluent mix-ture enclosed in a tube Because the projectile travels at supersonicspeeds oblique shock waves are formed at the nose of the projec-tile and they subsequently re ect from the side walls of the tubeand the projectile body Depending on the strength of these obliqueshocks ignition may occur after the initial shock or subsequentre ections Depending on the velocity of the projectile differentmodes of combustion and correspondingly different modes of op-eration of the system are possible When the velocity of the pro-jectile is greater than the CndashJ velocity an ODW may be stabilizedon the projectile resulting in the detonative mode of operation Aschematic of the ram accelerator operating in the detonative modeis shown in Fig 10 From the early days detonation has been dis-cussed as a means to attain very high velocities including escapevelocities37 In Ref 37 and several subsequent investigations thestructure of the detonative mode was described by the schematicdiagram shown in Fig 11 The projectile nose cone angle and thesound speed in the mixture selected were tailored to avoid combus-tion behind the initial shock wave but to provide initiation behindthe rst shock wave re ected from the tube wall Emphasis was alsoplaced on having the rst re ected shock wave strike the projec-tile body just aft of the shoulder37 If the system needed to be de-signed so precisely successful operation would indeed be dif cultin practical situations Further research has shown that the systemis far more stable and the particular scenario shown in Fig 11 ac-tually occurs just prior to system failure There were early reportsof successful attainment of superdetonative projectile velocities39

but the experimental emphasis shifted to the successful develop-ment of subdetonative modes of operation except at the Instituteof Saint-Louis (ISL) where research on the detonative mode hascontinued40 iexcl 42

Numerical simulations4344 and theoretical studies4546 based onsimpli ed analytical models of superdetonative ram acceleratorswere startedin the earlydays and have made asigni cant impactoverthe past decade These early studies con rmed the potential of ramaccelerators to accelerate ef ciently large masses to escape veloci-ties and low Earth orbit The steady-state simulations of Yungster47

highlighted the role of the projectile velocity (or Mach number)and showed that detonation could occur at different locations alongthe projectile body depending on the projectile Mach number (de- ned as the ratio of the velocity of the projectile to the speed ofsound in the premixed mixture) This work also raised the issue ofcombustion in the boundary layers along the projectile

The projectile velocity changes continuously in the ram acceler-ator and therefore steady-state simulations can at best be thoughtof as trying to characterize an instantaneous snapshot of the sys-tem The importance of the transient processes was rst highlightedin the work of Li et al48 This as well as subsequent work49 hasshown the importance of taking into account the time-dependent

Fig 12 Schematic showing the detailed structure of an oblique deto-nation and its relation to the detonation structures in ram accelerators

nature of the ow in the ram accelerator in describing phenomenasuch as unstarts and other system failures Computational studiesof the starting aspects of ram accelerators and the role of viscouseffects have also been conducted50 iexcl 52 Experimental efforts haveprimarily focused on the nondetonative mode except at ISL40 iexcl 42

Before discussing some of the experimental results and the dif cul-ties encountered in the practical development of ram accelerators inthe detonative mode two critical issues need to be discussed Oneissue is the actual structure of ODWs and the other is the stabilityof such detonation waves during transient operation

Structure and Stability of ODWsA key issue in the development of the ODWE and the ram accel-

erator is the structure and stability of ODWs Although the use ofODW was suggested as early as 195815 the stability of such waveswas implicitly assumed and questions about their detailed structurewere not considered Renewed interest in the late 1980s led to a the-oretical analysis of the supersonic ow of a combustible gas mixturepast a wedge53 Reference 53 indicated that for approach velocitiesroughly 25 or more larger than the CndashJ velocity of the reactantmixture ODWs can be stabilized over a useful range of wedge an-gles For wedge angles greater than a critical value the wave willdetach from the wedge and form an overdriven detonation wave ora more complex shock-induced combustion wave About this timethere were also numerical studies that showed that standing det-onation waves can be established on a wedge under certain owconditions5354

The early experimental efforts61314 were either focused on stabi-lizing a normal detonation wave in a duct or involved shock-inducedcombustion behind weak oblique shock waves Therefore it was notclear if ODWs could really be stabilized until some experiments inthe late 1980s on the diffraction and transmission of detonationsshowed that they could be5556

More recently the interest in ram accelerators has spawned alarge number of studies of the structure and stability of ODWs57 iexcl 61

The work of Li et al57 iexcl 59 showed that the basic structure of anoblique detonation generated by a wedge is more complex thanpreviously thought and consists of a nonreactive shock wave an in-duction region de agration waves and a detonation wave in whichthe pressure increase due to the energy release is closely coupledto the shock front This basic structure and a schematic of how itmight relate to the detonation waves in ram accelerator is shown inFig 12 These studies also showed that over a wide range of ow

KAILASANATH 1703

and mixture conditions the basic detonation structure is stable andvery resilient to disturbances in the ow In agreement with theory53

the simulations also showed the existence of a critical angle beyondwhich an attached stable detonation does not occur59 Other recentstudies have shown the same complex structure even with the inclu-sion of more complex chemical kinetic schemes6061 The essentialfeaturesof the complex ODW structure have also been con rmed byexperiments462 Thus a major conclusion from the recent researchis that the structure of an ODW can no longer be assumed to bea straight oblique shock wave followed closely by energy releaseand the complex structure can indeed be stabilizedon wedge-shapedobjects Implications of this work on propulsion applications mustbe considered

Although the schematic of Fig 12 shows how to relate the basicODW structure to that in ram accelerators the actual detonationwave structure in ram accelerators is likely to be more complexbecause of the expansion waves and multiple shock waves that aregenerated due to the geometric complexity of the system63 Theresults from a study of the detailed structure of detonations in ramaccelerators under a variety of ow conditions63 are summarizedin Fig 13 where the detonation wave structure is shown for three ow Mach numbers Recent experimental investigations have alsorevealed extremely complex detonation wave structures involvingmultiple shock waves under certain conditions64 In spite of theircomplexity all of these structures appear to be inherently relatedto the basic three-wave structure of oblique detonations shown inFig 12 However this appears to be an area of research that willcontinue for a while

a)

b)

c)

Fig 13 Detailed structure of detonations inram accelerators operatingat a) Mach 76 b) Mach 80 and c) Mach 84 (from Ref 63)

a) Case A Mach = 905

b) Case B Mach = 915

Fig 14 Visualization of the ow eld during the acceleration of theprojectile shown in Fig 5 from Mach 8 to 10 at two instants

Transient Nature of Detonation Waves in Ram AcceleratorsThe second important feature revealed by the recent studies is

the inherent transient nature of detonation waves in con gurationssuch as the ram accelerator When the concept of ram acceleratorsin the detonative mode was rst proposed it was thought that theleading shock wave (from the nose of the projectile) should re ectfrom the tube walls such that it impinges exactly at the shoulderof the projectile This will maximize the thrust and hence accel-eration However if the system were designed so precisely for onecondition its off-design performance could be unacceptable Fur-thermore simulations have shown that as the projectile acceleratesthe detonation wave structure tends to also move49 To illustrate thedynamics of detonations in the ram accelerator two photographsof the ow eld from a simulation (where the projectile acceleratedfrom Mach 8 to 10) are shown in Fig 14 In case A the projectileMach number is 905 and in case B it is 915 For both cases themain illustrations show the concentration of water vapor with thetwo inserts showing the pressure and temperature in a selected re-gion of the ow eld At the instant shown in Fig 14a (case A) thepressure and temperature behind the rst re ected shock wave arenot very high and combustion and energy release (as indicated bythe water vapor concentration) takes place only behind the secondre ected shock wave That is the detonation can be said to occurbehind the second re ected shock wave By the time the projectilehas accelerated to Mach 915 the leading shock wave has strength-ened raising the pressure and temperature behind the rst re ectedshock wave and causing the detonation to occur there

The behavior just discussedis seenthroughout the operationof theram accelerator in the detonative mode That is detonation waves inram accelerators are dynamic and move from shock wave to shockwave as the projectile acceleratesEventually the detonation wouldoccur just after the rst shock wave and the whole system will ceaseto operate because there will not be any positive thrust Furthermorethe detonation wave will run in front of the projectile creating a sit-uation similar to the classical unstart phenomena in ramjets Suchunstarts and detonations occurring after a different number of re- ections have also been noted in experiments42 Hence the shape ofthe projectile has to be chosen carefully so that the detonation canmove over the projectile body and still produce positive thrust overthe regime of interest A practical problem that has hindered thedevelopment of ram accelerators operating in the detonative modeis the change in the shape of the projectile due to material erosionburning and failure This is currently being overcome by the use ofsteel cowlings42

1704 KAILASANATH

PDEsThe other major current propulsion application involving detona-

tion waves is the PDE The basic concept behind a PDE is simpleDetonation is initiated repeatedly at either the closed or the openend of a detonation chamber that is lled with a premixed fuelndashairmixture Figure 15 (based on Ref 65) illustrates the concept forthe case when the detonation wave is initiated at the closed endAs shown in this schematic a one-dimensional planar detonation isinitiated near the closed end of the detonation chamber and travelsat the CndashJ velocity toward the open end A set of rarefaction wavestrail the detonation wave to reduce the velocity to zero at the closedend of the tube When the planar detonation wave exits the chamberanother set of expansion waves is generated that travels toward theclosed end These waves evacuate the burned detonation productsresulting in a cool empty chamber that is ready to be lled with afresh fuelndashair mixture The entire cycle repeats resulting in a pe-riodic high-pressure zone near the closed end of the chamber Theintegrated effectof this high pressure over the closed end (represent-ing a thrust wall) produces thrust This is of course a very idealizedpicture of the system Before discussing progress made in the re-cent years in developing a PDE a brief review of the background isprovided

The early work on the use of intermittent detonations has alreadybeen discussed In the late 1980s this concept of using intermittentor pulsed detonations was reexamined experimentally at the NavalPostgraduate School66 An ethylenendashoxygen detonation wave in asmall-diameter tube was used as a predetonator to initiate detona-tions in a larger tube containing an ethylenendashair mixture Periodicfuel injection within the naturally aspirated tube resulted in an in-termittent frequency of 25 Hz Speci c impulse estimated usingthe pressure time history and the amount of fuel consumed rangedfrom 1000 to 1400 s The velocities of the observed waves (lessthan 1 kms) are signi cantly below the CndashJ detonation wave ve-locities for the reported mixtures indicating that a fully developeddetonation wave was not formed

The rst purely computational study of the PDE reported in theopen literature appears to be the work of Cambier and Adelman67

The system simulated consisted of a 50-cm-long main tube attachedto a 43-cm-long diverging nozzle Quasi-one-dimensional simula-tions using the Euler equations for unsteady ow were carried outwith a total variation diminishing scheme and multistep nite ratekinetics A detonation was initiated at the closed end of a tube lledwith a premixed stoichiometric hydrogenndashair mixture at 3 atmOverall performance calculated by integrating the instantaneousthrust and the fuel ow rates gave a speci c impulse Isp of about6500 s and a range of operating frequencies up to 667 Hz Issues ofignition energies and transition to detonation were not consideredand the dif culties with attaining rapid mixing of fuel and air wasavoided by using a premixed fuelndashair mixture These assumptionsare still typical of those used in more recent simulations

The rst look at the interactions between the ow inside the cham-ber and that outside appears to be the work of Eidelman et al7

who simulated a cylindrical detonation chamber (about 15-cm diam

Fig 15 Schematic of an idealized PDE showing various stages duringone cycle of operation (adapted from Ref 65)

Fig 16 Generic PDE device used for somecomputational studieswheredetonation is initiated near the open end (from Ref 69)

and 15 cm long) with a small converging nozzle In their two-dimensional simulations a self-similarsolution for a planar detona-tion was imposed near the open end and made to travel toward theclosed end The re ection of the wave from the closed end createda higher pressure at the wall than during initiation near the closedend Hence this mode of initiation was considered to be better Af-ter the detonation left the chamber the pressure in the chamber wasobserved to fall below the ambient resulting in ingestion of outsideair In subsequent works86869 the same numerical techniques andmodels were used to investigate a slightly different con gurationwhere the nozzle at the aft end was removed and an inlet was in-troduced near the closed end This con guration has become oneof the popular systems for two-dimensional (and axisymmetric) nu-merical simulations Again detonation was assumed to be initiatedat the open end and traveled toward the closed end In this con gu-ration when the pressure falls below atmospheric in the detonationchamber air is ingested through the inlets in the front even duringstatic operations Thus a valveless self-aspiratingoperation is pos-sible These simulations also introduced the discussion of open-endinitiation vs closed-end initiation and their relative advantages Thecurrent status of this issue will be discussed later

Many of the early efforts on intermittent or PDEs have been re-viewed in Ref 7 and 8 where a link is also made between thesedetonation engines and pulse jet engines (which do not use detona-tive combustion) In addition Ref 8 provides a detailed summaryof the numerical investigations conducted by the authorsrsquo group ona generic PDE device shown in Fig 16 A characteristic feature ofthis device is that the detonation wave is initiated at the open endand travels toward a thrust wall at the closed end An advantage ofthis approach is that it is valveless and self-aspirating Air is en-trained into the chamber through inlets near the closed end whenthe pressure falls below atmospheric pressure during part of the cy-cle Unsteady Euler simulations were used to study the operationof the device from static (M = 0) to supersonic (M = 2) conditionsAs stated in Ref 8 a proof of principle experimental demonstra-tion of the PDE mode of operation where detonation is initiated atthe open end still needed to be done Furthermore issues such asdetonation initiation and structure and fuelndashair injection and mixingalso needed to be studied

Bussing and Pappas65 provided a detailed description of thebasic operation of an ideal zed PDE They also reported someone-dimensional studies of PDEs burning hydrogenndashoxygen andhydrogenndashair mixtures The detonation wave was initiated at theclosedend in this engine using a high-temperature and high-pressureregion An advantage of this design is the con nement provided atthe closed end which should enable a more rapid initiation andestablishment of a detonation Both airbreathing and rocket modeof operation of this valved design was discussed Lynch et al70

presented a computational study of an axisymmetric PDE with astraight inlet very similar to that of Eidelman and Grossmann8 andEidelman et al69 Various inlet lengths from 1 to 12 cm and ightMach numbers of 08 and 20 were studied The results were similarto those reported earlier except that some additional details of thedetonation wave front and the air induction system were resolved inthese studies In a related work Lynch and Edelman71 showed thatreductions in scavenging time (and hence overall cycle time) canbe achieved by suitably shaping the inlet

Bussing et al72 compared open-end initiation and closed-end ini-tiation in a simple tube PDE geometry Equivalent thrust productionand fuel ef ciency were observed for the two types of initiation

KAILASANATH 1705

Although a higher head-end pressure was attained during open-endinitiation the pressure continuously dropped for that case unlike theclosed-end initiation where a constant pressure region occurred fora while before reduction of the pressure to the ambient value Morerecently an experimental study has con rmed the differences in theobserved pressure pro les and it has also been concluded that themaximum level of impulse attained is independent of the locationof the detonation initiator73

In the past few years there has been a substantial growth in theappearance of papers related to the PDE and complete sessionshave been devoted to the topic at annual AIAAASMESAEASEEjoint propulsion conferences Because of space considerations onlyselected papers from these years are brie y discussed here Sterlinget al74 conducted one-dimensional numerical investigations of aself-aspiratingPDE operating for multiple cyclesA key observationfrom their studies was that only a portion of the detonation tubecan be lled with fresh charge under self-aspirating operation Inspite of this limitation they reported a speci c impulse of 5151 sfor a hydrogenndashair system However they concluded that the idealperformance of such an engine is ldquonear those of other hydrogen-fueledair breathing enginesrdquo74

Several basic shock tube experiments related to the developmentof a hydrogen-fueled PDE were described by Hinkey et al75 Theseinclude the measurement of detonation wave velocities and de a-gration to detonation transition (DDT) lengths for a range of equiv-alence ratios As expected the DDT lengths were found to be toolarge for practical applications even in hydrogenndashoxygen mixturesTherefore traditional techniques such as the use of a Schelkin spiral(see Ref 76) were tried to reduce the transition length A factor of2ndash4 reduction was observed over the range of equivalence ratios in-vestigated This work75 highlights the dif culties involved in obtain-ing a fully developed detonation in a short distance (as assumed inthe conceptual studies and numerical simulations discussed earlier)and questions if the detonations reported in previous experimentalinvestigations161766 were really fully transitioned from de agra-tions The speci c impulses measured in their experiments (about240 s for hydrogenndashoxygen and 1200 s for hydrogenndashair) were alsostated to be in agreement with their analysis

An interesting PDE concept presented by Bussing77 was that of arotary-valved multiple-pulsed detonation engine Here several det-onation chambers are coupled to an air inlet and fuel source usinga rotary valve as suggested in one of the early papers of Nichollset al16 The rotary valves allow the lling of some detonation cham-bers while others are detonating or exhausting Other papers includepresentation of an overall PDE performance model78 and the exper-imental characterization of the detonation properties of some fuelsfor use in PDEs79

The importance of adequately mixing the fuel and oxidizer washighlighted by the experimental investigations of Stanley et al80

who obtained very low sub-CndashJ velocities when injecting the fueland oxidizer at different times and not taking extra effort to en-sure that they were well mixed The use of turbulence producingdevices appeared to improve signi cantly the mixing and the at-tainment of higher velocities but also resulted in signi cant thrustlosses This study also showed that raising the initial pressure wasbene cial The computational studies of Eidelman et al81 showedthat the PDE engine could operate even for a range of transitionaldetonation regimes that produced nonplanar or not fully developeddetonations Aarnio et al82 discussed two failure modes DDT tran-sition failure and premature ignition In both cases the system con-tinued to operate though there would be loss of thrust In Ref 82detailed discussions of the pressure and thrust histories during a5-Hz 20-cycle operation are also provided (Fig 17) The cycle tocycle variations of peak pressure have been attributed to the lowsampling frequency used82 Both a direct measurement as well asthe integration of the pressure histories were used to estimate the av-erage Isp from hydrogenndashair systems to be between 1116 and 1333 sThe effect of the predetonator volume on the reduction of speci cimpulse was also shown to be a critical issue The effects of inletsand nozzles were not considered in this study

A conceptual design of a multitube PDE with a single air inletduct was discussed by Pegg et al83 Time-dependent computational

Fig 17 Pressure history at one location from a PDE operating at 5 Hzfor 20 cycles (from Ref 82)

uid dynamics analysis indicated that the inlet isolaterdiffusercon-cept would work and did not allow enough time for the formationof destabilizing hammer shocks Performance analysis includingcomponent ef ciencies reported in the paper indicate that opera-tional frequencies of the order of 75ndash100 Hz are required for theMach 12ndash3 ights considered Experimental data indicating mul-ticycle operation at 100 Hz with a rich ethylenendashoxygen mixturewere presented by Sterling et al84 DDT enhancement devices andpredetonators were used in the experiments In spite of this det-onation waves did not occur all of the time and when detonationwaves failed the tube became very hot and testing was suspendedTherefore the maximum testing time was reported to be 05 s Inspite of this limitation the two studies8384 taken together high-light the possibility of PDEs burning hydrocarbon fuels for realisticmissions

Five different nozzle shapes and their effect on performance werestudied computationally by Cambier and Tegner85 Their results in-dicate that the presence of a nozzle can affect the performance ofthe PDE by increasing the thrust delivery during the ignition phaseThe bell-shaped nozzles appeared to give higher performance thanshapes with positive curvature Their results also showed that noz-zles also affected the ow dynamics and hence the timing of thevarious phases of the engine cycle In an experimental study of aconical exhaust nozzle the authors of Ref 86 noted improved per-formance (based on higher velocities which were still sub-CndashJ) butdid not see any effect on the blowdown process More recently87

the effectof various nozzle shapes including converging divergingand straight have been reexamined computationally The converg-ing sections of nozzles introduced shock wave re ections whereasdiverging sections generated a negative thrust for a portion of thecycle due to overexpansion In spite of these limitations the overallconclusion of these studies was that ldquonozzles can drastically in-crease ef ciency of the PDEsrdquo87 There is also some experimentalevidence suggesting that propulsive performances can be increasedby the addition of a nozzle73 However factors such as the effect ofthe nozzles on detonation transmission and the detailed dynamicsof the ow have not yet been elucidated This appears to be an areathat needs further investigation

Other recent efforts have focused on demonstrating the opera-tion of single- and multitube combustors coupled to an inlet using arotary-valvemechanism88 The valve servesto both meter the air owinto the combustor and to isolate the inlet from the high pressuresproduced during the detonation cycle Utilizing multiple combus-tors that ll and detonate out of phase allows the continual use of thein owing air Firing rates of up to 12 Hz per combustor were demon-strated for a hydrogen-fueled system Several other issues related tothe practical development of PDEs have been discussed by Bussinget al89 According to them three of the more important engineer-ing issues are the development of a valve-injection subsystem anignition system and a low-volume predetonator

The use of PDEs for rockets has been revived recently90 Therocket mode of operation is very similar to the description providedearlier of the airbreathing engine with ignition at the closed endexcept that the oxidizer also needs to be injected into the system pe-riodically An advantage of PDEs for rockets would be their higherpower density thus enabling the development of more compact

1706 KAILASANATH

rockets Reference 90 shows the pressure traces from a pulsed det-onation rocket engine operating at 145 Hz on a hydrogenndashoxygenmixture

More recently a two-tube rotary-valved PDE with ight-sizecomponents was operated at 40 Hz per combustor for 30 s withan ethylenendashair mixture91 In addition throttling of the device al-lowed operation at three thrust levels during a 10-s demonstrationexperiment Contrary to some earlier estimates the thermal loadswere signi cant and it was not possible to operate for more than3 s without water cooling It is not clear if this was because ofperiodic detonation failures and consequent de agrative modes ofcombustion as reported in some earlier studies

The work reported in the open literature and discussed earlier hasfocused on gaseous fuels However for volume-limited propulsionapplications PDEs operating on liquid fuels need to be demon-strated In a recent study Brophy et al92 reported on experimentsusing JP-10oxygen and JP-10air aerosols The fuelndashoxygen mix-ture was successfully detonated to obtain an engine operating at5 Hz but the tests with the fuelndashair mixture were not successful

Computational studies continue to make progress in the study ofthe PDE8793 iexcl 97 The effect of nozzle shapes and incomplete transi-tion to CndashJ detonations have been studied by Eidelman and Yang87

They nd that the cycle ef ciency will be virtually the same whethera CndashJ detonation wave is initiated instantaneously or if the CndashJ valueis attainedonly at the end of the detonation chamber Primarilybasedon this study87 they come to an interesting conclusion that ldquoinitia-tion energies that are required for PDE operations can be small andcomparable to energies required for initiation of (other) combustionprocessesrdquo However this is yet to be substantiated experimentallySekar et al93 reexamine open- and closed-end initiation and thebasic self-aspirating con guration of Eidelman8 using more recentcomputational tools Injection and mixing issues that have gener-ally been ignored in previous numerical studies are also beginningto be addressed94 Both front-wall and lateral-wall injections werestudied and the time required for fuel injection was estimated Itis suggested that it may be possible to ll only a portion of thechamber during high-frequency operations Recent computationalstudies also suggest that preconditioning the fuelndashair mixture us-ing a shock wave can signi cantly reduce the DDT distance95 Theperformance estimates mentioned throughout the discussion of thePDE show signi cant variation A systematic review of various per-formance estimates and possible reasons for the observed variationhave been presented9697 These numerical studies suggest that thetime history of the back pressure is a crucial parameter affecting theestimated performance

Other Detonation EnginesThere have also been several other interesting conceptual studies

on the applications of detonations to engines Cambier et al98 haveproposed a pulsed-detonation wave augmentation device as part ofa hybrid engine for a single-stage-to-orbit airbreathing hypersonicvehicle Here the PDE is used for both thrust generation as well asfor mixingcombustion augmentation Two speci c con gurationswere investigated numerically

The concept of a detonation internal combustion engine was pro-posed by Loth and Loth99 and Loth100 Here a conventional pistoncylinder con guration is modi ed to incorporate a separate deto-nation chamber which is isolated by a valve from the compressionchamber during fuel injection and discharges tangentially into a vor-texchamber formedby the piston and cylinder at top deadcenterTherapid detonative combustion is followed by rapid dilution throughmixing with air in the vortex chamber to reduce the formation ofNOx and unburned hydrocarbons and to achieve an overall leanmixture The vortex chamber was also designed to store a portion ofthe detonation waversquos kinetic energy Overall thermal ef cienciessomewhat better than constant volume combustion were postulatedunder ideal conditions

ConclusionsDetonation waves have been explored extensively for propulsion

applications because of their inherent theoretical advantage overde agrative combustion However practical developments to date

have been of idealized systems or laboratory-scale devices or bothThe basic ideas behind most of the current systems being investi-gated have been known for a while but still many of the details needto be understood Advances in computations and experimental di-agnostics appear to be poised to make practical propulsion devicesbased on detonation waves a reality

Two major systems currently under development are the ram ac-celerator and the PDE In both cases the theoretical and computa-tional studies have been far more encouraging than the experimentsHowever in both cases actual demonstrations of devices workingon gaseous fuels have taken place In the case of the ram accelera-tor fuelndashoxygenndashdiluent mixtures have been used and the primaryproblem appears to be heat transfer and related material failuresThe use of gaseous fuels is not a major issue here but new materialsfor the projectiles need to be explored

In the case of the PDE most of the successful operations havebeen with fuelndashoxygen mixtures There still has not been a directcomparison between detailed experimental data and related theoret-ical and computational results Furthermore most of the PDE andrelated work to date have focused on gaseous mixtures PDEs op-erating on multiphase mixtures need to be emphasized because itwould be more practical to use liquid fuels for most of the propul-sion applicationsproposed for these systemsThere are several othergeneral issues that need to be resolved Clearly the evacuation ofexhaust gases is important both from the point of view of controllingthe cycle time and for preventing autoignition and mixing with freshgases The heat transfer and noise from PDEs have also not beenreported in any detail Other issues include inletndashcombustor andcombustorndashnozzle interactions and scalability of the entire device

AcknowledgmentsThis work has been sponsored by the Of ce of Naval Research

through the Mechanics and Energy Conversion Division and theUS Naval Research Laboratory

References1Fickett W and Davis W C Detonation Univ of California Press

Berkeley CA 19792Hoffmann N ldquoReaction Propulsion by Intermittent Detonative Com-

bustionrdquo German Ministry of Supply AI152365 Volkenrode Translation1940 (cited in Ref 16)

3Roy M ldquoPropulsion par Statoreacteur a Detonationrdquo Comptes-Rendusde lrsquoAcademie des Sciences Vol 222 1946 pp 31 32

4Dabora E K and Broda J C ldquoStanding Normal Detonations andOblique Detonations for Propulsionrdquo AIAA Paper 93-2325 July 1993

5Dabora E K ldquoStatus of Gaseous Detonation Waves and Their Rolein Propulsionrdquo Fall Technical Meeting of the Eastern States Section ofthe Combustion Institute Combustion Inst Pittsburgh PA 1994 pp 11ndash

186Rubins P M and Bauer R C ldquoReview of Shock-Induced Supersonic

Combustion Research and Hypersonic Applicationsrdquo Journal of Propulsionand Power Vol 10 No 5 1994 pp 593ndash601

7Eidelman S Grossmann W and Lottati I ldquoReview of PropulsionApplications and Numerical Simulations of the Pulsed Detonation EngineConceptrdquo Journal of Propulsion and Power Vol 7 No 6 1991 pp 857ndash

865 also AIAA Paper 89-2446 July 19898Eidelman S and Grossmann W ldquoPulsed Detonation Engine Experi-

mental and Theoretical Reviewrdquo AIAA Paper 92-3168 July 19929Reingold L ldquoRecherches sur les combustions permanentes apportees

aux foyers a circulation interne supersoniquerdquo ONERA TN2 Dept of En-ergy and Propulsion Study 728-E Paris March 1950

10Bitondo D and Bollay W ldquoPreliminary Performance Analysis ofthe Pulse-Detonation-Jet Engine Systemrdquo Aerophysics Development CorpRept ADC-102-1 April 1952

11Gross R A ldquoExploratory Study of Combustion in Supersonic FlowrdquoUS Air Force Of ce of Scienti c Research TN 59-587 Washington DCJune 1959

12Gross R A and Chinitz W ldquoA Study of Supersonic CombustionrdquoJournal of Aerospace Science Vol 27 No 7 1960 pp 517ndash525

13Nicholls J A Dabora E K and Gealer R L ldquoStudies in Con-nection with Stabilized Gaseous Detonation Wavesrdquo Seventh Symposium(International) on Combustion Combustion Inst Pittsburgh PA 1959 pp766ndash772

14Nicholls J A and Dabora E K ldquoRecent Results on Standing Detona-tion Wavesrdquo Eighth Symposium (International) on Combustion CombustionInst Pittsburgh PA 1962 pp 644ndash655

KAILASANATH 1707

15Dunlap R Brehm R L and Nicholls J A ldquoA Preliminary Studyof the Application of Steady-State Detonative Combustion to a ReactionEnginerdquo Jet Propulsion Vol 28 No 7 1958 pp 451ndash456

16Nicholls J A Wilkinson H R and Morrison R B ldquoIntermittentDetonation as a Thrust-Producing Mechanismrdquo Jet Propulsion Vol 27No 5 1957 pp 534ndash541

17Krzycki L J ldquoPerformance Characteristics of an Intermittent Detona-tion Devicerdquo US Naval Ordnance Test Station US Naval Weapons Rept7655 China Lake CA 1962

18Nicholls J A Cullen R E and Ragland K W ldquoFeasibility Studiesof a Rotating Detonation Wave Rocket Motorrdquo Journal of Spacecraft andRockets Vol 3 No 6 1966 pp 893ndash898

19Adamson T C and Olsson G R ldquoPerformance Analysis of a RotatingDetonation Wave Rocket Enginerdquo Astronautica Acta Vol 13 No 4 1967pp 405ndash415

20Shen P I and Adamson T C ldquoTheoretical Analysis of a RotatingTwo-Phase Detonation in Liquid Rocket Motorsrdquo Astronautica Acta Vol17 No 4 1972 pp 715ndash728

21Clayton R M and Rogero R S ldquoExperimental Measurements on aRotating Detonation-Like Wave Observed During Liquid Rocket ResonantCombustionrdquo Jet Propulsion Lab TR 32-788 California Inst of Technol-ogy Pasadena CA Aug 1965

22Back L H ldquoApplication of Blast Wave Theory to Explosive Propul-sionrdquo Acta Astronautica Vol 2 1975 pp 391ndash407

23Varsi G Back L H and Kim K ldquoBlast Wave in a Nozzle for Propul-sion Applicationsrdquo Acta Astronautica Vol 3 1976 pp 141ndash156

24Kim K Varsi G and Back LH ldquoBlast Wave Analysis for DetonationPropulsionrdquo AIAA Journal Vol 10 No 10 1977 pp 1500ndash1502

25Back L H Dowler W L and Varsi G ldquoDetonation PropulsionExperiments and Theoryrdquo AIAA Journal Vol 21 No 10 1983 pp 1418ndash

142726Kantrowitz A ldquoPropulsion to Orbit by Ground Based Lasersrdquo Astro-

nautics and Aeronautics Vol 10 No 5 1972 p 7427Hyde R A ldquoOne-D Modelling of a Two-Pulse LSD Thrusterrdquo Pro-

ceedings of the 1986 SDIODARPA Workshop on Laser Propulsion editedby J T Kare Lawrence Livermore National Labs 1987 pp 79ndash88

28Emery M H Kailasanath K Oran E S and Gardner J H ldquoCompu-tational Studies of Laser Supported Detonationsrdquo Proceedings of the 1987SDIO Workshop on Laser Propulsion edited by J T Kare Lawrence Liv-ermore National Labs 1990 pp 43ndash55

29Emery M H Kailasanath K and Oran E S ldquoNumerical Simu-lations of Laser Supported Detonationsrdquo 1988 SDIO Workshop on LaserPropulsion Lawrence Livermore National Labs April 1988

30Carrier G F Fendell F McGregor D Cook S and Vazirani MldquoLaser-Initiated Conical Detonation Wave for Supersonic CombustionrdquoAIAA Paper 91-0578 Jan 1991

31Fendell F E Mitchell J McGregor R Magiawala K and Shef eldM ldquoLaser-Initiated Conical Detonation Wave for Supersonic CombustionIIrdquo AIAA Paper 92-0088 Jan 1992

32Carrier G F Fendell F E and Chou M S ldquoLaser-Initiated ConicalDetonation Wave for Supersonic Combustion IIIrdquo AIAA Paper 92-3247July 1992

33Menees G P Adelman H G Cambier J-L and Bowles J V ldquoWaveCombustors for Trans-Atmospheric Vehiclesrdquo Journal of Propulsion andPower Vol 8 No 3 1992 pp 709ndash713

34Cambier J L Adelman H and Menees G ldquoNumerical Simulationsof an Oblique Detonation Wave Enginerdquo AIAA Paper 88-0063 1988

35Atamanchuk T and Sislian J ldquoOn- and Off-design Performance Anal-ysis of Hypersonic Detonation Wave Ramjetsrdquo AIAA Paper 90-2473 1990

36Kuznetsov M M Neyland V J and Sayapin G N ldquoEf ciency In-vestigation of Scramjet with Detonation and Shockless Supersonic Combus-tionrdquo Utchenye Zapisky TsAGI Vol 23 No 2 1992 pp 30ndash37

37Bruckner A P and Hertzberg A ldquoRam Accelerator Direct LaunchSystem for Space Cargordquo International Astronautical Federation Paper 87-211 1987

38Hertzberg A Bruckner A P and Bogdanoff D W ldquoRam Accel-erator A New Chemical Method for Accelerating Projectiles to UltrahighVelocitiesrdquo AIAA Journal Vol 26 No 2 1988 pp 195ndash203

39Kull A Burnham E Knowlen C Bruckner A P and HertzbergA ldquoExperimental Studies of the Superdetonative Ram Accelerator ModesrdquoAIAA Paper 89-2632 1989

40Patz G Seiler F Smeets G and Srulijes J ldquoStatus of ISLrsquos RA-MAC30 with Fin Guided Projectiles Accelerated in a Smooth Borerdquo Pro-ceedings of the Second International Workshop on Ram Accelerators Univof Washington Seattle WA 1995

41Seiler F Patz G Smeets G and Srulijes J ldquoGasdynamic Limitsof Ignition and Combustion of a Gas Mixture in ISLrsquos RAMAC30 ScramAcceleratorrdquo Proceedings of the Second International Workshop on RamAccelerators Univ of Washington Seattle WA 1995

42Seiler F Patz G Smeets G and Srulijes J ldquoIn uence of ProjectileMaterial and Gas Composition on Superdetonative Combustion in ISLrsquosRAMAC30rdquo AIAA Paper 98-3445 July 1998

43Brackett D C and Bogdanoff D W ldquoComputational Investigation ofOblique Detonation Ramjet-in-Tube Conceptsrdquo Journal of Propulsion andPower Vol 5 No 3 1989 pp 276ndash281

44Yungster S Eberhardt S and Bruckner A P ldquoNumerical Simula-tion of Shock-Induced Combustion Generated by High-Speed Projectiles inDetonable Gas Mixturesrdquo AIAA Paper 89-0673 1989

45Rom J and Kivity Y ldquoAccelerating Projectiles Up to 12 kms Utiliz-ing the Continuous Detonation Propulsion Methodrdquo AIAA Paper 88-29691988

46Humphrey J W ldquoParametric Study of an ODW Scramaccelerator forHypersonic Test Facilitiesrdquo AIAA Paper 90-2470 1990

47Yungster S ldquoNavierndashStokes Simulation of the Supersonic CombustionFlow eld in a Ram Acceleratorrdquo AIAA Paper 91-1916 1991

48Li C Kailasanath K Oran E S Boris J P Paper and LandsbergA M ldquoNumerical Simulations of Transient Flows in Ram AcceleratorsrdquoAIAA Paper 93-1916 1993

49Li C Kailasanath K Oran E S Landsberg A M and Boris J PldquoDynamics of Oblique Detonations in Ram Acceleratorsrdquo Shock Waves Vol5 June 1995 pp 97ndash101

50Yungster S and Radhakrishnan K ldquoComputational Study of FlowEstablishment in a Ram Acceleratorrdquo AIAA Paper 98-2489 July 1995

51Choi J Y Jeung I-S and Yoon Y ldquoNumerical Study of ScramAccelerator Starting Characteristicsrdquo AIAA Journal Vol 36 No 6 1998pp 1029ndash1038

52Bezgin L Ganzhelo A Gouskov O and Kopchenov V ldquoSome Nu-merical Investigation Results of Shock-Induced Combustionrdquo Proceedingsof the AIAA 8th International Space Planes and Hypersonic Systems andTechnologies Conference AIAA Reston VA 1998 pp 72ndash82

53Pratt D W Humphrey J W and Glenn D E ldquoMorphology of Stand-ing Oblique Detonation Wavesrdquo Journal of Propulsion and Power Vol 7No 5 1991

54Fujiwara T Matsuo A and Nomoto H ldquoTwo-Dimensional Detona-tion Supported by a Blunt Body or a Wedgerdquo AIAA Paper 88-0089 Jan1988

55Fan B C Sichel M and Kauffman C W ldquoAnalysis of ObliqueShock-Detonation Wave Interactions in the Supersonic Flow of a Com-bustible Mediumrdquo AIAA Paper 88-0441 Jan 1988

56Dabora E K Desbordes D Guerraud C and Wagner H GldquoOblique Detonations at Hypersonic Velocitiesrdquo Dynamics of Detonationsand Explosions Detonations Vol 133 Progress in Astronautics and Aero-nautics AIAA Washington DC 1990 pp 187ndash201

57Li C Kailasanath K and Oran E S ldquoStability of Oblique Detona-tions in Ram Acceleratorsrdquo AIAA Paper 92-0089 Jan 1992

58Li C Kailasanath K and Oran E S ldquoStructure of Reaction WavesBehind Oblique Shocksrdquo Dynamic Aspects of Detonation Vol 153 Progressin Astronautics and Aeronautics AIAA Washington DC 1993 pp 231ndash

24059Li C Kailasanath K and Oran E S ldquoDetonation Structures Behind

Oblique Shocksrdquo Physics of Fluids Vol 6 April 1994 pp 1600ndash161160Da Silva L F and Deshaies B ldquoNumerical Study of Some Ignition

Regimes of Combustible Supersonic Flows over a Wedgerdquo Proceedings ofSecond International Workshop on Ram Accelerators Paper 15 Univ ofWashington Seattle WA 1995

61Thaker A A and Chelliah H K ldquoNumerical Prediction of ObliqueDetonation Wave Structures Using Detailed and Reduced Reaction Mecha-nismsrdquo Combustion Theory and Modelling Vol 1 No 4 1997 pp 347ndash376

62Desbordes D Guerraud C and Hamada L ldquoSupersonic H2-AirCombustions Behind Oblique Shock Wavesrdquo 14th International Colloquiumon Dynamics of Explosions and Reactive Systems Paper D28 Univ ofCoimbra Portugal Aug 1993

63Li C Kailasanath K and Oran E S ldquoDetonation Structures Gener-ated by Multiple Shocks on Ram-Accelerator Projectilesrdquo Combustion andFlame Vol 108 1997 pp 173ndash186

64Viguier C and Desbordes D ldquoConditions of Onset and Stabilizationof Oblique Detonation Wavesrdquo Advances in Experimentation and Compu-tation of Detonations edited by G D Roy S H Frolov K Kailasanathand N Smirnov ENAS Moscow 1998 pp 40 41

65Bussing Tand Pappas G ldquoIntroduction toPulseDetonation EnginesrdquoAIAA Paper 94-0263 1994

66Helman D Shreeve R P and Eidelman S ldquoDetonation Pulse En-ginerdquo AIAA Paper 86-1683 June 1986

67Cambier J L and Adelman H G ldquoPreliminary Numerical Simu-lations of a Pulsed Detonation Wave Enginerdquo AIAA Paper 88-2960 July1988

68Eidelman S Grossmann W and Lottati I ldquoAir-Breathing PulsedDetonationEngine Concept A Numerical Studyrdquo AIAA Paper 90-2420 July1990

69Eidelman S Grossmann W and Lottati I ldquoComputational Analysisof Pulsed Detonation Engines and Applicationsrdquo AIAA Paper 90-0460 Jan1990

KAILASANATH 1701

detonation propulsion technology can be considered for producingboth large velocity changes as a main propulsion system or for gen-erating small velocity changes such as in attitudecontrol propulsionAlthough this can be considered to be one of the more successfulapplications of detonation to propulsion one must remember thatin this application the details of the detonation process itself ap-pear to be secondary and the working uid is a nondetonable gasmixture

Laser-Supported Detonations (LSDs)Another intriguing concept proposed in the 1970s was the use of

lasers to propel objects to low Earth orbit26 The basic idea behindthis propulsion concept is to use lasers to generate detonations anda high kinetic energy propulsive jet moving away from the objectbeing propelled The idea was not pursued very far due to the limitedlaser powers available then and the obvious atmospheric transmis-sion problems However with the planned development of largerlasers under the Strategic Defense Initiative program the conceptwas explored in some detail (for example see Refs 27ndash29) in thelate 1980s In principle a high controllable exhaust velocity can beachieved by a two-step process In the rst step the desired amountof propellant is formed as a gas layer near the surface and in thesecond step the laser energy is directly deposited in this layer The rst layer can be generated either by transpirationof a liquid througha porous surface or by a laser prepulse in a two-pulse system27 Ina two-pulse system a very low-intensity long-duration (on the or-der of seconds) laser prepulse is used to ablate enough gas fromthe surface of the solidliquid propellant to form a buffer gas layerA second high-intensity short-duration laser pulse irradiates thegas layer depositing its energy in the gas layer just before reachingthe propellant surface A shock wave then propagates out throughthe gas layer back toward the laser heating and accelerating thegas layer What is effectively a laser-supported detonation (LSD)is formed by the shock wave followed by the recombination zonebehind it28 In theory a high speci c impulse ( laquo 1000 s) can begenerated with simple propellants such as ice However the realproblems of building a large enough laser and safely transmittingenough energy to the desired location remain and to the best ofthe authorrsquos knowledge LSDs have not been conclusively demon-strated

Another concept using lasers to initiate detonations was proposedby Carrier et al30 Fendell et al31 and Carrier et al32 They con-ducted theoretical and experimental feasibility studies of a super-sonic combustor based on a stabilized conically con gured obliquedetonation wave (ODW) The stabilized detonation wave is formedby the interaction of a train of spherical detonation waves each di-rectly initiated by a brief localized deposition of energy from a veryrapidly repeated pulsed laser The laser is focused on a xed site inthe combustor where there is a steady uniform supersonic streamof detonable gases Initiation of an individual spherical detonationwave by a single laser pulse was achieved32 but further results havenot been reported

ODWEsWith the renewed emphasis on hypersonic ight vehicles in the

1980s extensive investigations were conducted of supersonic com-bustion concepts However these studies will not be discussed hereunless they speci callyinvoke detonations Various versionsof theseconcepts have been called by special names such as detonation-driven ramjets (dramjets)5 but these essentially involve stabilizedODWs and can be traced to the original concept proposed in the1950s by Dunlap et al15 Basically if a premixed fuelndashair mixture ows at a velocity greater than the CndashJ velocity then a normal det-onation wave cannot be stabilized The hope was that under theseconditions an ODW can be stabilized on a wedge-shaped objectsuch that only the normal component of the ow becomes sonic orsubsonic That is most of the ow in the combustor will remainsupersonic A schematic of the dramjet is shown in Fig 8 (fromRef 5) Here fuel is injected into the incoming airstream mixesrapidly with it and is ignited behind the centerbody of a ramjetwhich acts as a stabilizer holding the oblique wave It is similar toa conventional ramjet except for the close coupling between the re-

Fig 8 Schematic of a dramjet (from Ref 5)

Fig 9 Schematic of shock structure for scramjet engine and an ODWE(from Ref 33)

action waves and the leading shock waves resulting in a detonationwave The performance of such a device has been calculated andshown5 to be less than that of an ideal ramjet over a range of Machnumbers from 5 to 30 Issues such as the presence of inlet shocksand the ability to mix rapidly enough to obtain a detonable mixturewere not addressed in this paper5

A representation closer to scramjet engines is shown in Fig 9(from Ref 33) Here the inlet shock system is shown but the linkbetween the inlet shocks and the ODW is left unclear Once againit is assumed that fuelndashair mixing occurs rapidly enough and thatthe oblique shock wave is strong enough so that chemical reactionsoccur close behind it and are coupled to it Analysis of a conceptualvehicle using such an ODWE showed that it has better performancethan an equivalent scramjet for Mach numbers over 15 The mainadvantage arises from the reduced length of detonation engines In arelatedwork the issueof mixing inhomogeneities and their effectonthe detonation wave were addressed numerically34 There have alsobeen severalother studies examining the feasibilityand performanceof detonation wave ramjets3536 All of these concepts essentiallyrely on rapid ignition and combustion occurring behind an obliqueshock wave and the coupling between the two processes resultingin a stable structure A difference between these more recent worksand the earlier ones is the realization that in many situations shock-induced combustion does not lead to detonations However none ofthese studies seem to have led to the actual building and testing ofan ODWE

Ram AcceleratorsIf the solid body anchoring the oblique shock or detonation is

free to move the integral of pressure forces around the body canresult in a net force causing the body to accelerate or decelerateThis is essentially the concept behind the ram accelerator a propul-sion concept proposed3738 in the late 1980s In a ram accelerator

1702 KAILASANATH

Fig 10 Schematic of a ram accelerator in the detonative mode

Fig 11 Early concept of the detonative mode (from Ref 37)

a projectile that resembles the centerbody of a ramjet travels at asupersonic speed through a premixed fuelndashoxidizerndashdiluent mix-ture enclosed in a tube Because the projectile travels at supersonicspeeds oblique shock waves are formed at the nose of the projec-tile and they subsequently re ect from the side walls of the tubeand the projectile body Depending on the strength of these obliqueshocks ignition may occur after the initial shock or subsequentre ections Depending on the velocity of the projectile differentmodes of combustion and correspondingly different modes of op-eration of the system are possible When the velocity of the pro-jectile is greater than the CndashJ velocity an ODW may be stabilizedon the projectile resulting in the detonative mode of operation Aschematic of the ram accelerator operating in the detonative modeis shown in Fig 10 From the early days detonation has been dis-cussed as a means to attain very high velocities including escapevelocities37 In Ref 37 and several subsequent investigations thestructure of the detonative mode was described by the schematicdiagram shown in Fig 11 The projectile nose cone angle and thesound speed in the mixture selected were tailored to avoid combus-tion behind the initial shock wave but to provide initiation behindthe rst shock wave re ected from the tube wall Emphasis was alsoplaced on having the rst re ected shock wave strike the projec-tile body just aft of the shoulder37 If the system needed to be de-signed so precisely successful operation would indeed be dif cultin practical situations Further research has shown that the systemis far more stable and the particular scenario shown in Fig 11 ac-tually occurs just prior to system failure There were early reportsof successful attainment of superdetonative projectile velocities39

but the experimental emphasis shifted to the successful develop-ment of subdetonative modes of operation except at the Instituteof Saint-Louis (ISL) where research on the detonative mode hascontinued40 iexcl 42

Numerical simulations4344 and theoretical studies4546 based onsimpli ed analytical models of superdetonative ram acceleratorswere startedin the earlydays and have made asigni cant impactoverthe past decade These early studies con rmed the potential of ramaccelerators to accelerate ef ciently large masses to escape veloci-ties and low Earth orbit The steady-state simulations of Yungster47

highlighted the role of the projectile velocity (or Mach number)and showed that detonation could occur at different locations alongthe projectile body depending on the projectile Mach number (de- ned as the ratio of the velocity of the projectile to the speed ofsound in the premixed mixture) This work also raised the issue ofcombustion in the boundary layers along the projectile

The projectile velocity changes continuously in the ram acceler-ator and therefore steady-state simulations can at best be thoughtof as trying to characterize an instantaneous snapshot of the sys-tem The importance of the transient processes was rst highlightedin the work of Li et al48 This as well as subsequent work49 hasshown the importance of taking into account the time-dependent

Fig 12 Schematic showing the detailed structure of an oblique deto-nation and its relation to the detonation structures in ram accelerators

nature of the ow in the ram accelerator in describing phenomenasuch as unstarts and other system failures Computational studiesof the starting aspects of ram accelerators and the role of viscouseffects have also been conducted50 iexcl 52 Experimental efforts haveprimarily focused on the nondetonative mode except at ISL40 iexcl 42

Before discussing some of the experimental results and the dif cul-ties encountered in the practical development of ram accelerators inthe detonative mode two critical issues need to be discussed Oneissue is the actual structure of ODWs and the other is the stabilityof such detonation waves during transient operation

Structure and Stability of ODWsA key issue in the development of the ODWE and the ram accel-

erator is the structure and stability of ODWs Although the use ofODW was suggested as early as 195815 the stability of such waveswas implicitly assumed and questions about their detailed structurewere not considered Renewed interest in the late 1980s led to a the-oretical analysis of the supersonic ow of a combustible gas mixturepast a wedge53 Reference 53 indicated that for approach velocitiesroughly 25 or more larger than the CndashJ velocity of the reactantmixture ODWs can be stabilized over a useful range of wedge an-gles For wedge angles greater than a critical value the wave willdetach from the wedge and form an overdriven detonation wave ora more complex shock-induced combustion wave About this timethere were also numerical studies that showed that standing det-onation waves can be established on a wedge under certain owconditions5354

The early experimental efforts61314 were either focused on stabi-lizing a normal detonation wave in a duct or involved shock-inducedcombustion behind weak oblique shock waves Therefore it was notclear if ODWs could really be stabilized until some experiments inthe late 1980s on the diffraction and transmission of detonationsshowed that they could be5556

More recently the interest in ram accelerators has spawned alarge number of studies of the structure and stability of ODWs57 iexcl 61

The work of Li et al57 iexcl 59 showed that the basic structure of anoblique detonation generated by a wedge is more complex thanpreviously thought and consists of a nonreactive shock wave an in-duction region de agration waves and a detonation wave in whichthe pressure increase due to the energy release is closely coupledto the shock front This basic structure and a schematic of how itmight relate to the detonation waves in ram accelerator is shown inFig 12 These studies also showed that over a wide range of ow

KAILASANATH 1703

and mixture conditions the basic detonation structure is stable andvery resilient to disturbances in the ow In agreement with theory53

the simulations also showed the existence of a critical angle beyondwhich an attached stable detonation does not occur59 Other recentstudies have shown the same complex structure even with the inclu-sion of more complex chemical kinetic schemes6061 The essentialfeaturesof the complex ODW structure have also been con rmed byexperiments462 Thus a major conclusion from the recent researchis that the structure of an ODW can no longer be assumed to bea straight oblique shock wave followed closely by energy releaseand the complex structure can indeed be stabilizedon wedge-shapedobjects Implications of this work on propulsion applications mustbe considered

Although the schematic of Fig 12 shows how to relate the basicODW structure to that in ram accelerators the actual detonationwave structure in ram accelerators is likely to be more complexbecause of the expansion waves and multiple shock waves that aregenerated due to the geometric complexity of the system63 Theresults from a study of the detailed structure of detonations in ramaccelerators under a variety of ow conditions63 are summarizedin Fig 13 where the detonation wave structure is shown for three ow Mach numbers Recent experimental investigations have alsorevealed extremely complex detonation wave structures involvingmultiple shock waves under certain conditions64 In spite of theircomplexity all of these structures appear to be inherently relatedto the basic three-wave structure of oblique detonations shown inFig 12 However this appears to be an area of research that willcontinue for a while

a)

b)

c)

Fig 13 Detailed structure of detonations inram accelerators operatingat a) Mach 76 b) Mach 80 and c) Mach 84 (from Ref 63)

a) Case A Mach = 905

b) Case B Mach = 915

Fig 14 Visualization of the ow eld during the acceleration of theprojectile shown in Fig 5 from Mach 8 to 10 at two instants

Transient Nature of Detonation Waves in Ram AcceleratorsThe second important feature revealed by the recent studies is

the inherent transient nature of detonation waves in con gurationssuch as the ram accelerator When the concept of ram acceleratorsin the detonative mode was rst proposed it was thought that theleading shock wave (from the nose of the projectile) should re ectfrom the tube walls such that it impinges exactly at the shoulderof the projectile This will maximize the thrust and hence accel-eration However if the system were designed so precisely for onecondition its off-design performance could be unacceptable Fur-thermore simulations have shown that as the projectile acceleratesthe detonation wave structure tends to also move49 To illustrate thedynamics of detonations in the ram accelerator two photographsof the ow eld from a simulation (where the projectile acceleratedfrom Mach 8 to 10) are shown in Fig 14 In case A the projectileMach number is 905 and in case B it is 915 For both cases themain illustrations show the concentration of water vapor with thetwo inserts showing the pressure and temperature in a selected re-gion of the ow eld At the instant shown in Fig 14a (case A) thepressure and temperature behind the rst re ected shock wave arenot very high and combustion and energy release (as indicated bythe water vapor concentration) takes place only behind the secondre ected shock wave That is the detonation can be said to occurbehind the second re ected shock wave By the time the projectilehas accelerated to Mach 915 the leading shock wave has strength-ened raising the pressure and temperature behind the rst re ectedshock wave and causing the detonation to occur there

The behavior just discussedis seenthroughout the operationof theram accelerator in the detonative mode That is detonation waves inram accelerators are dynamic and move from shock wave to shockwave as the projectile acceleratesEventually the detonation wouldoccur just after the rst shock wave and the whole system will ceaseto operate because there will not be any positive thrust Furthermorethe detonation wave will run in front of the projectile creating a sit-uation similar to the classical unstart phenomena in ramjets Suchunstarts and detonations occurring after a different number of re- ections have also been noted in experiments42 Hence the shape ofthe projectile has to be chosen carefully so that the detonation canmove over the projectile body and still produce positive thrust overthe regime of interest A practical problem that has hindered thedevelopment of ram accelerators operating in the detonative modeis the change in the shape of the projectile due to material erosionburning and failure This is currently being overcome by the use ofsteel cowlings42

1704 KAILASANATH

PDEsThe other major current propulsion application involving detona-

tion waves is the PDE The basic concept behind a PDE is simpleDetonation is initiated repeatedly at either the closed or the openend of a detonation chamber that is lled with a premixed fuelndashairmixture Figure 15 (based on Ref 65) illustrates the concept forthe case when the detonation wave is initiated at the closed endAs shown in this schematic a one-dimensional planar detonation isinitiated near the closed end of the detonation chamber and travelsat the CndashJ velocity toward the open end A set of rarefaction wavestrail the detonation wave to reduce the velocity to zero at the closedend of the tube When the planar detonation wave exits the chamberanother set of expansion waves is generated that travels toward theclosed end These waves evacuate the burned detonation productsresulting in a cool empty chamber that is ready to be lled with afresh fuelndashair mixture The entire cycle repeats resulting in a pe-riodic high-pressure zone near the closed end of the chamber Theintegrated effectof this high pressure over the closed end (represent-ing a thrust wall) produces thrust This is of course a very idealizedpicture of the system Before discussing progress made in the re-cent years in developing a PDE a brief review of the background isprovided

The early work on the use of intermittent detonations has alreadybeen discussed In the late 1980s this concept of using intermittentor pulsed detonations was reexamined experimentally at the NavalPostgraduate School66 An ethylenendashoxygen detonation wave in asmall-diameter tube was used as a predetonator to initiate detona-tions in a larger tube containing an ethylenendashair mixture Periodicfuel injection within the naturally aspirated tube resulted in an in-termittent frequency of 25 Hz Speci c impulse estimated usingthe pressure time history and the amount of fuel consumed rangedfrom 1000 to 1400 s The velocities of the observed waves (lessthan 1 kms) are signi cantly below the CndashJ detonation wave ve-locities for the reported mixtures indicating that a fully developeddetonation wave was not formed

The rst purely computational study of the PDE reported in theopen literature appears to be the work of Cambier and Adelman67

The system simulated consisted of a 50-cm-long main tube attachedto a 43-cm-long diverging nozzle Quasi-one-dimensional simula-tions using the Euler equations for unsteady ow were carried outwith a total variation diminishing scheme and multistep nite ratekinetics A detonation was initiated at the closed end of a tube lledwith a premixed stoichiometric hydrogenndashair mixture at 3 atmOverall performance calculated by integrating the instantaneousthrust and the fuel ow rates gave a speci c impulse Isp of about6500 s and a range of operating frequencies up to 667 Hz Issues ofignition energies and transition to detonation were not consideredand the dif culties with attaining rapid mixing of fuel and air wasavoided by using a premixed fuelndashair mixture These assumptionsare still typical of those used in more recent simulations

The rst look at the interactions between the ow inside the cham-ber and that outside appears to be the work of Eidelman et al7

who simulated a cylindrical detonation chamber (about 15-cm diam

Fig 15 Schematic of an idealized PDE showing various stages duringone cycle of operation (adapted from Ref 65)

Fig 16 Generic PDE device used for somecomputational studieswheredetonation is initiated near the open end (from Ref 69)

and 15 cm long) with a small converging nozzle In their two-dimensional simulations a self-similarsolution for a planar detona-tion was imposed near the open end and made to travel toward theclosed end The re ection of the wave from the closed end createda higher pressure at the wall than during initiation near the closedend Hence this mode of initiation was considered to be better Af-ter the detonation left the chamber the pressure in the chamber wasobserved to fall below the ambient resulting in ingestion of outsideair In subsequent works86869 the same numerical techniques andmodels were used to investigate a slightly different con gurationwhere the nozzle at the aft end was removed and an inlet was in-troduced near the closed end This con guration has become oneof the popular systems for two-dimensional (and axisymmetric) nu-merical simulations Again detonation was assumed to be initiatedat the open end and traveled toward the closed end In this con gu-ration when the pressure falls below atmospheric in the detonationchamber air is ingested through the inlets in the front even duringstatic operations Thus a valveless self-aspiratingoperation is pos-sible These simulations also introduced the discussion of open-endinitiation vs closed-end initiation and their relative advantages Thecurrent status of this issue will be discussed later

Many of the early efforts on intermittent or PDEs have been re-viewed in Ref 7 and 8 where a link is also made between thesedetonation engines and pulse jet engines (which do not use detona-tive combustion) In addition Ref 8 provides a detailed summaryof the numerical investigations conducted by the authorsrsquo group ona generic PDE device shown in Fig 16 A characteristic feature ofthis device is that the detonation wave is initiated at the open endand travels toward a thrust wall at the closed end An advantage ofthis approach is that it is valveless and self-aspirating Air is en-trained into the chamber through inlets near the closed end whenthe pressure falls below atmospheric pressure during part of the cy-cle Unsteady Euler simulations were used to study the operationof the device from static (M = 0) to supersonic (M = 2) conditionsAs stated in Ref 8 a proof of principle experimental demonstra-tion of the PDE mode of operation where detonation is initiated atthe open end still needed to be done Furthermore issues such asdetonation initiation and structure and fuelndashair injection and mixingalso needed to be studied

Bussing and Pappas65 provided a detailed description of thebasic operation of an ideal zed PDE They also reported someone-dimensional studies of PDEs burning hydrogenndashoxygen andhydrogenndashair mixtures The detonation wave was initiated at theclosedend in this engine using a high-temperature and high-pressureregion An advantage of this design is the con nement provided atthe closed end which should enable a more rapid initiation andestablishment of a detonation Both airbreathing and rocket modeof operation of this valved design was discussed Lynch et al70

presented a computational study of an axisymmetric PDE with astraight inlet very similar to that of Eidelman and Grossmann8 andEidelman et al69 Various inlet lengths from 1 to 12 cm and ightMach numbers of 08 and 20 were studied The results were similarto those reported earlier except that some additional details of thedetonation wave front and the air induction system were resolved inthese studies In a related work Lynch and Edelman71 showed thatreductions in scavenging time (and hence overall cycle time) canbe achieved by suitably shaping the inlet

Bussing et al72 compared open-end initiation and closed-end ini-tiation in a simple tube PDE geometry Equivalent thrust productionand fuel ef ciency were observed for the two types of initiation

KAILASANATH 1705

Although a higher head-end pressure was attained during open-endinitiation the pressure continuously dropped for that case unlike theclosed-end initiation where a constant pressure region occurred fora while before reduction of the pressure to the ambient value Morerecently an experimental study has con rmed the differences in theobserved pressure pro les and it has also been concluded that themaximum level of impulse attained is independent of the locationof the detonation initiator73

In the past few years there has been a substantial growth in theappearance of papers related to the PDE and complete sessionshave been devoted to the topic at annual AIAAASMESAEASEEjoint propulsion conferences Because of space considerations onlyselected papers from these years are brie y discussed here Sterlinget al74 conducted one-dimensional numerical investigations of aself-aspiratingPDE operating for multiple cyclesA key observationfrom their studies was that only a portion of the detonation tubecan be lled with fresh charge under self-aspirating operation Inspite of this limitation they reported a speci c impulse of 5151 sfor a hydrogenndashair system However they concluded that the idealperformance of such an engine is ldquonear those of other hydrogen-fueledair breathing enginesrdquo74

Several basic shock tube experiments related to the developmentof a hydrogen-fueled PDE were described by Hinkey et al75 Theseinclude the measurement of detonation wave velocities and de a-gration to detonation transition (DDT) lengths for a range of equiv-alence ratios As expected the DDT lengths were found to be toolarge for practical applications even in hydrogenndashoxygen mixturesTherefore traditional techniques such as the use of a Schelkin spiral(see Ref 76) were tried to reduce the transition length A factor of2ndash4 reduction was observed over the range of equivalence ratios in-vestigated This work75 highlights the dif culties involved in obtain-ing a fully developed detonation in a short distance (as assumed inthe conceptual studies and numerical simulations discussed earlier)and questions if the detonations reported in previous experimentalinvestigations161766 were really fully transitioned from de agra-tions The speci c impulses measured in their experiments (about240 s for hydrogenndashoxygen and 1200 s for hydrogenndashair) were alsostated to be in agreement with their analysis

An interesting PDE concept presented by Bussing77 was that of arotary-valved multiple-pulsed detonation engine Here several det-onation chambers are coupled to an air inlet and fuel source usinga rotary valve as suggested in one of the early papers of Nichollset al16 The rotary valves allow the lling of some detonation cham-bers while others are detonating or exhausting Other papers includepresentation of an overall PDE performance model78 and the exper-imental characterization of the detonation properties of some fuelsfor use in PDEs79

The importance of adequately mixing the fuel and oxidizer washighlighted by the experimental investigations of Stanley et al80

who obtained very low sub-CndashJ velocities when injecting the fueland oxidizer at different times and not taking extra effort to en-sure that they were well mixed The use of turbulence producingdevices appeared to improve signi cantly the mixing and the at-tainment of higher velocities but also resulted in signi cant thrustlosses This study also showed that raising the initial pressure wasbene cial The computational studies of Eidelman et al81 showedthat the PDE engine could operate even for a range of transitionaldetonation regimes that produced nonplanar or not fully developeddetonations Aarnio et al82 discussed two failure modes DDT tran-sition failure and premature ignition In both cases the system con-tinued to operate though there would be loss of thrust In Ref 82detailed discussions of the pressure and thrust histories during a5-Hz 20-cycle operation are also provided (Fig 17) The cycle tocycle variations of peak pressure have been attributed to the lowsampling frequency used82 Both a direct measurement as well asthe integration of the pressure histories were used to estimate the av-erage Isp from hydrogenndashair systems to be between 1116 and 1333 sThe effect of the predetonator volume on the reduction of speci cimpulse was also shown to be a critical issue The effects of inletsand nozzles were not considered in this study

A conceptual design of a multitube PDE with a single air inletduct was discussed by Pegg et al83 Time-dependent computational

Fig 17 Pressure history at one location from a PDE operating at 5 Hzfor 20 cycles (from Ref 82)

uid dynamics analysis indicated that the inlet isolaterdiffusercon-cept would work and did not allow enough time for the formationof destabilizing hammer shocks Performance analysis includingcomponent ef ciencies reported in the paper indicate that opera-tional frequencies of the order of 75ndash100 Hz are required for theMach 12ndash3 ights considered Experimental data indicating mul-ticycle operation at 100 Hz with a rich ethylenendashoxygen mixturewere presented by Sterling et al84 DDT enhancement devices andpredetonators were used in the experiments In spite of this det-onation waves did not occur all of the time and when detonationwaves failed the tube became very hot and testing was suspendedTherefore the maximum testing time was reported to be 05 s Inspite of this limitation the two studies8384 taken together high-light the possibility of PDEs burning hydrocarbon fuels for realisticmissions

Five different nozzle shapes and their effect on performance werestudied computationally by Cambier and Tegner85 Their results in-dicate that the presence of a nozzle can affect the performance ofthe PDE by increasing the thrust delivery during the ignition phaseThe bell-shaped nozzles appeared to give higher performance thanshapes with positive curvature Their results also showed that noz-zles also affected the ow dynamics and hence the timing of thevarious phases of the engine cycle In an experimental study of aconical exhaust nozzle the authors of Ref 86 noted improved per-formance (based on higher velocities which were still sub-CndashJ) butdid not see any effect on the blowdown process More recently87

the effectof various nozzle shapes including converging divergingand straight have been reexamined computationally The converg-ing sections of nozzles introduced shock wave re ections whereasdiverging sections generated a negative thrust for a portion of thecycle due to overexpansion In spite of these limitations the overallconclusion of these studies was that ldquonozzles can drastically in-crease ef ciency of the PDEsrdquo87 There is also some experimentalevidence suggesting that propulsive performances can be increasedby the addition of a nozzle73 However factors such as the effect ofthe nozzles on detonation transmission and the detailed dynamicsof the ow have not yet been elucidated This appears to be an areathat needs further investigation

Other recent efforts have focused on demonstrating the opera-tion of single- and multitube combustors coupled to an inlet using arotary-valvemechanism88 The valve servesto both meter the air owinto the combustor and to isolate the inlet from the high pressuresproduced during the detonation cycle Utilizing multiple combus-tors that ll and detonate out of phase allows the continual use of thein owing air Firing rates of up to 12 Hz per combustor were demon-strated for a hydrogen-fueled system Several other issues related tothe practical development of PDEs have been discussed by Bussinget al89 According to them three of the more important engineer-ing issues are the development of a valve-injection subsystem anignition system and a low-volume predetonator

The use of PDEs for rockets has been revived recently90 Therocket mode of operation is very similar to the description providedearlier of the airbreathing engine with ignition at the closed endexcept that the oxidizer also needs to be injected into the system pe-riodically An advantage of PDEs for rockets would be their higherpower density thus enabling the development of more compact

1706 KAILASANATH

rockets Reference 90 shows the pressure traces from a pulsed det-onation rocket engine operating at 145 Hz on a hydrogenndashoxygenmixture

More recently a two-tube rotary-valved PDE with ight-sizecomponents was operated at 40 Hz per combustor for 30 s withan ethylenendashair mixture91 In addition throttling of the device al-lowed operation at three thrust levels during a 10-s demonstrationexperiment Contrary to some earlier estimates the thermal loadswere signi cant and it was not possible to operate for more than3 s without water cooling It is not clear if this was because ofperiodic detonation failures and consequent de agrative modes ofcombustion as reported in some earlier studies

The work reported in the open literature and discussed earlier hasfocused on gaseous fuels However for volume-limited propulsionapplications PDEs operating on liquid fuels need to be demon-strated In a recent study Brophy et al92 reported on experimentsusing JP-10oxygen and JP-10air aerosols The fuelndashoxygen mix-ture was successfully detonated to obtain an engine operating at5 Hz but the tests with the fuelndashair mixture were not successful

Computational studies continue to make progress in the study ofthe PDE8793 iexcl 97 The effect of nozzle shapes and incomplete transi-tion to CndashJ detonations have been studied by Eidelman and Yang87

They nd that the cycle ef ciency will be virtually the same whethera CndashJ detonation wave is initiated instantaneously or if the CndashJ valueis attainedonly at the end of the detonation chamber Primarilybasedon this study87 they come to an interesting conclusion that ldquoinitia-tion energies that are required for PDE operations can be small andcomparable to energies required for initiation of (other) combustionprocessesrdquo However this is yet to be substantiated experimentallySekar et al93 reexamine open- and closed-end initiation and thebasic self-aspirating con guration of Eidelman8 using more recentcomputational tools Injection and mixing issues that have gener-ally been ignored in previous numerical studies are also beginningto be addressed94 Both front-wall and lateral-wall injections werestudied and the time required for fuel injection was estimated Itis suggested that it may be possible to ll only a portion of thechamber during high-frequency operations Recent computationalstudies also suggest that preconditioning the fuelndashair mixture us-ing a shock wave can signi cantly reduce the DDT distance95 Theperformance estimates mentioned throughout the discussion of thePDE show signi cant variation A systematic review of various per-formance estimates and possible reasons for the observed variationhave been presented9697 These numerical studies suggest that thetime history of the back pressure is a crucial parameter affecting theestimated performance

Other Detonation EnginesThere have also been several other interesting conceptual studies

on the applications of detonations to engines Cambier et al98 haveproposed a pulsed-detonation wave augmentation device as part ofa hybrid engine for a single-stage-to-orbit airbreathing hypersonicvehicle Here the PDE is used for both thrust generation as well asfor mixingcombustion augmentation Two speci c con gurationswere investigated numerically

The concept of a detonation internal combustion engine was pro-posed by Loth and Loth99 and Loth100 Here a conventional pistoncylinder con guration is modi ed to incorporate a separate deto-nation chamber which is isolated by a valve from the compressionchamber during fuel injection and discharges tangentially into a vor-texchamber formedby the piston and cylinder at top deadcenterTherapid detonative combustion is followed by rapid dilution throughmixing with air in the vortex chamber to reduce the formation ofNOx and unburned hydrocarbons and to achieve an overall leanmixture The vortex chamber was also designed to store a portion ofthe detonation waversquos kinetic energy Overall thermal ef cienciessomewhat better than constant volume combustion were postulatedunder ideal conditions

ConclusionsDetonation waves have been explored extensively for propulsion

applications because of their inherent theoretical advantage overde agrative combustion However practical developments to date

have been of idealized systems or laboratory-scale devices or bothThe basic ideas behind most of the current systems being investi-gated have been known for a while but still many of the details needto be understood Advances in computations and experimental di-agnostics appear to be poised to make practical propulsion devicesbased on detonation waves a reality

Two major systems currently under development are the ram ac-celerator and the PDE In both cases the theoretical and computa-tional studies have been far more encouraging than the experimentsHowever in both cases actual demonstrations of devices workingon gaseous fuels have taken place In the case of the ram accelera-tor fuelndashoxygenndashdiluent mixtures have been used and the primaryproblem appears to be heat transfer and related material failuresThe use of gaseous fuels is not a major issue here but new materialsfor the projectiles need to be explored

In the case of the PDE most of the successful operations havebeen with fuelndashoxygen mixtures There still has not been a directcomparison between detailed experimental data and related theoret-ical and computational results Furthermore most of the PDE andrelated work to date have focused on gaseous mixtures PDEs op-erating on multiphase mixtures need to be emphasized because itwould be more practical to use liquid fuels for most of the propul-sion applicationsproposed for these systemsThere are several othergeneral issues that need to be resolved Clearly the evacuation ofexhaust gases is important both from the point of view of controllingthe cycle time and for preventing autoignition and mixing with freshgases The heat transfer and noise from PDEs have also not beenreported in any detail Other issues include inletndashcombustor andcombustorndashnozzle interactions and scalability of the entire device

AcknowledgmentsThis work has been sponsored by the Of ce of Naval Research

through the Mechanics and Energy Conversion Division and theUS Naval Research Laboratory

References1Fickett W and Davis W C Detonation Univ of California Press

Berkeley CA 19792Hoffmann N ldquoReaction Propulsion by Intermittent Detonative Com-

bustionrdquo German Ministry of Supply AI152365 Volkenrode Translation1940 (cited in Ref 16)

3Roy M ldquoPropulsion par Statoreacteur a Detonationrdquo Comptes-Rendusde lrsquoAcademie des Sciences Vol 222 1946 pp 31 32

4Dabora E K and Broda J C ldquoStanding Normal Detonations andOblique Detonations for Propulsionrdquo AIAA Paper 93-2325 July 1993

5Dabora E K ldquoStatus of Gaseous Detonation Waves and Their Rolein Propulsionrdquo Fall Technical Meeting of the Eastern States Section ofthe Combustion Institute Combustion Inst Pittsburgh PA 1994 pp 11ndash

186Rubins P M and Bauer R C ldquoReview of Shock-Induced Supersonic

Combustion Research and Hypersonic Applicationsrdquo Journal of Propulsionand Power Vol 10 No 5 1994 pp 593ndash601

7Eidelman S Grossmann W and Lottati I ldquoReview of PropulsionApplications and Numerical Simulations of the Pulsed Detonation EngineConceptrdquo Journal of Propulsion and Power Vol 7 No 6 1991 pp 857ndash

865 also AIAA Paper 89-2446 July 19898Eidelman S and Grossmann W ldquoPulsed Detonation Engine Experi-

mental and Theoretical Reviewrdquo AIAA Paper 92-3168 July 19929Reingold L ldquoRecherches sur les combustions permanentes apportees

aux foyers a circulation interne supersoniquerdquo ONERA TN2 Dept of En-ergy and Propulsion Study 728-E Paris March 1950

10Bitondo D and Bollay W ldquoPreliminary Performance Analysis ofthe Pulse-Detonation-Jet Engine Systemrdquo Aerophysics Development CorpRept ADC-102-1 April 1952

11Gross R A ldquoExploratory Study of Combustion in Supersonic FlowrdquoUS Air Force Of ce of Scienti c Research TN 59-587 Washington DCJune 1959

12Gross R A and Chinitz W ldquoA Study of Supersonic CombustionrdquoJournal of Aerospace Science Vol 27 No 7 1960 pp 517ndash525

13Nicholls J A Dabora E K and Gealer R L ldquoStudies in Con-nection with Stabilized Gaseous Detonation Wavesrdquo Seventh Symposium(International) on Combustion Combustion Inst Pittsburgh PA 1959 pp766ndash772

14Nicholls J A and Dabora E K ldquoRecent Results on Standing Detona-tion Wavesrdquo Eighth Symposium (International) on Combustion CombustionInst Pittsburgh PA 1962 pp 644ndash655

KAILASANATH 1707

15Dunlap R Brehm R L and Nicholls J A ldquoA Preliminary Studyof the Application of Steady-State Detonative Combustion to a ReactionEnginerdquo Jet Propulsion Vol 28 No 7 1958 pp 451ndash456

16Nicholls J A Wilkinson H R and Morrison R B ldquoIntermittentDetonation as a Thrust-Producing Mechanismrdquo Jet Propulsion Vol 27No 5 1957 pp 534ndash541

17Krzycki L J ldquoPerformance Characteristics of an Intermittent Detona-tion Devicerdquo US Naval Ordnance Test Station US Naval Weapons Rept7655 China Lake CA 1962

18Nicholls J A Cullen R E and Ragland K W ldquoFeasibility Studiesof a Rotating Detonation Wave Rocket Motorrdquo Journal of Spacecraft andRockets Vol 3 No 6 1966 pp 893ndash898

19Adamson T C and Olsson G R ldquoPerformance Analysis of a RotatingDetonation Wave Rocket Enginerdquo Astronautica Acta Vol 13 No 4 1967pp 405ndash415

20Shen P I and Adamson T C ldquoTheoretical Analysis of a RotatingTwo-Phase Detonation in Liquid Rocket Motorsrdquo Astronautica Acta Vol17 No 4 1972 pp 715ndash728

21Clayton R M and Rogero R S ldquoExperimental Measurements on aRotating Detonation-Like Wave Observed During Liquid Rocket ResonantCombustionrdquo Jet Propulsion Lab TR 32-788 California Inst of Technol-ogy Pasadena CA Aug 1965

22Back L H ldquoApplication of Blast Wave Theory to Explosive Propul-sionrdquo Acta Astronautica Vol 2 1975 pp 391ndash407

23Varsi G Back L H and Kim K ldquoBlast Wave in a Nozzle for Propul-sion Applicationsrdquo Acta Astronautica Vol 3 1976 pp 141ndash156

24Kim K Varsi G and Back LH ldquoBlast Wave Analysis for DetonationPropulsionrdquo AIAA Journal Vol 10 No 10 1977 pp 1500ndash1502

25Back L H Dowler W L and Varsi G ldquoDetonation PropulsionExperiments and Theoryrdquo AIAA Journal Vol 21 No 10 1983 pp 1418ndash

142726Kantrowitz A ldquoPropulsion to Orbit by Ground Based Lasersrdquo Astro-

nautics and Aeronautics Vol 10 No 5 1972 p 7427Hyde R A ldquoOne-D Modelling of a Two-Pulse LSD Thrusterrdquo Pro-

ceedings of the 1986 SDIODARPA Workshop on Laser Propulsion editedby J T Kare Lawrence Livermore National Labs 1987 pp 79ndash88

28Emery M H Kailasanath K Oran E S and Gardner J H ldquoCompu-tational Studies of Laser Supported Detonationsrdquo Proceedings of the 1987SDIO Workshop on Laser Propulsion edited by J T Kare Lawrence Liv-ermore National Labs 1990 pp 43ndash55

29Emery M H Kailasanath K and Oran E S ldquoNumerical Simu-lations of Laser Supported Detonationsrdquo 1988 SDIO Workshop on LaserPropulsion Lawrence Livermore National Labs April 1988

30Carrier G F Fendell F McGregor D Cook S and Vazirani MldquoLaser-Initiated Conical Detonation Wave for Supersonic CombustionrdquoAIAA Paper 91-0578 Jan 1991

31Fendell F E Mitchell J McGregor R Magiawala K and Shef eldM ldquoLaser-Initiated Conical Detonation Wave for Supersonic CombustionIIrdquo AIAA Paper 92-0088 Jan 1992

32Carrier G F Fendell F E and Chou M S ldquoLaser-Initiated ConicalDetonation Wave for Supersonic Combustion IIIrdquo AIAA Paper 92-3247July 1992

33Menees G P Adelman H G Cambier J-L and Bowles J V ldquoWaveCombustors for Trans-Atmospheric Vehiclesrdquo Journal of Propulsion andPower Vol 8 No 3 1992 pp 709ndash713

34Cambier J L Adelman H and Menees G ldquoNumerical Simulationsof an Oblique Detonation Wave Enginerdquo AIAA Paper 88-0063 1988

35Atamanchuk T and Sislian J ldquoOn- and Off-design Performance Anal-ysis of Hypersonic Detonation Wave Ramjetsrdquo AIAA Paper 90-2473 1990

36Kuznetsov M M Neyland V J and Sayapin G N ldquoEf ciency In-vestigation of Scramjet with Detonation and Shockless Supersonic Combus-tionrdquo Utchenye Zapisky TsAGI Vol 23 No 2 1992 pp 30ndash37

37Bruckner A P and Hertzberg A ldquoRam Accelerator Direct LaunchSystem for Space Cargordquo International Astronautical Federation Paper 87-211 1987

38Hertzberg A Bruckner A P and Bogdanoff D W ldquoRam Accel-erator A New Chemical Method for Accelerating Projectiles to UltrahighVelocitiesrdquo AIAA Journal Vol 26 No 2 1988 pp 195ndash203

39Kull A Burnham E Knowlen C Bruckner A P and HertzbergA ldquoExperimental Studies of the Superdetonative Ram Accelerator ModesrdquoAIAA Paper 89-2632 1989

40Patz G Seiler F Smeets G and Srulijes J ldquoStatus of ISLrsquos RA-MAC30 with Fin Guided Projectiles Accelerated in a Smooth Borerdquo Pro-ceedings of the Second International Workshop on Ram Accelerators Univof Washington Seattle WA 1995

41Seiler F Patz G Smeets G and Srulijes J ldquoGasdynamic Limitsof Ignition and Combustion of a Gas Mixture in ISLrsquos RAMAC30 ScramAcceleratorrdquo Proceedings of the Second International Workshop on RamAccelerators Univ of Washington Seattle WA 1995

42Seiler F Patz G Smeets G and Srulijes J ldquoIn uence of ProjectileMaterial and Gas Composition on Superdetonative Combustion in ISLrsquosRAMAC30rdquo AIAA Paper 98-3445 July 1998

43Brackett D C and Bogdanoff D W ldquoComputational Investigation ofOblique Detonation Ramjet-in-Tube Conceptsrdquo Journal of Propulsion andPower Vol 5 No 3 1989 pp 276ndash281

44Yungster S Eberhardt S and Bruckner A P ldquoNumerical Simula-tion of Shock-Induced Combustion Generated by High-Speed Projectiles inDetonable Gas Mixturesrdquo AIAA Paper 89-0673 1989

45Rom J and Kivity Y ldquoAccelerating Projectiles Up to 12 kms Utiliz-ing the Continuous Detonation Propulsion Methodrdquo AIAA Paper 88-29691988

46Humphrey J W ldquoParametric Study of an ODW Scramaccelerator forHypersonic Test Facilitiesrdquo AIAA Paper 90-2470 1990

47Yungster S ldquoNavierndashStokes Simulation of the Supersonic CombustionFlow eld in a Ram Acceleratorrdquo AIAA Paper 91-1916 1991

48Li C Kailasanath K Oran E S Boris J P Paper and LandsbergA M ldquoNumerical Simulations of Transient Flows in Ram AcceleratorsrdquoAIAA Paper 93-1916 1993

49Li C Kailasanath K Oran E S Landsberg A M and Boris J PldquoDynamics of Oblique Detonations in Ram Acceleratorsrdquo Shock Waves Vol5 June 1995 pp 97ndash101

50Yungster S and Radhakrishnan K ldquoComputational Study of FlowEstablishment in a Ram Acceleratorrdquo AIAA Paper 98-2489 July 1995

51Choi J Y Jeung I-S and Yoon Y ldquoNumerical Study of ScramAccelerator Starting Characteristicsrdquo AIAA Journal Vol 36 No 6 1998pp 1029ndash1038

52Bezgin L Ganzhelo A Gouskov O and Kopchenov V ldquoSome Nu-merical Investigation Results of Shock-Induced Combustionrdquo Proceedingsof the AIAA 8th International Space Planes and Hypersonic Systems andTechnologies Conference AIAA Reston VA 1998 pp 72ndash82

53Pratt D W Humphrey J W and Glenn D E ldquoMorphology of Stand-ing Oblique Detonation Wavesrdquo Journal of Propulsion and Power Vol 7No 5 1991

54Fujiwara T Matsuo A and Nomoto H ldquoTwo-Dimensional Detona-tion Supported by a Blunt Body or a Wedgerdquo AIAA Paper 88-0089 Jan1988

55Fan B C Sichel M and Kauffman C W ldquoAnalysis of ObliqueShock-Detonation Wave Interactions in the Supersonic Flow of a Com-bustible Mediumrdquo AIAA Paper 88-0441 Jan 1988

56Dabora E K Desbordes D Guerraud C and Wagner H GldquoOblique Detonations at Hypersonic Velocitiesrdquo Dynamics of Detonationsand Explosions Detonations Vol 133 Progress in Astronautics and Aero-nautics AIAA Washington DC 1990 pp 187ndash201

57Li C Kailasanath K and Oran E S ldquoStability of Oblique Detona-tions in Ram Acceleratorsrdquo AIAA Paper 92-0089 Jan 1992

58Li C Kailasanath K and Oran E S ldquoStructure of Reaction WavesBehind Oblique Shocksrdquo Dynamic Aspects of Detonation Vol 153 Progressin Astronautics and Aeronautics AIAA Washington DC 1993 pp 231ndash

24059Li C Kailasanath K and Oran E S ldquoDetonation Structures Behind

Oblique Shocksrdquo Physics of Fluids Vol 6 April 1994 pp 1600ndash161160Da Silva L F and Deshaies B ldquoNumerical Study of Some Ignition

Regimes of Combustible Supersonic Flows over a Wedgerdquo Proceedings ofSecond International Workshop on Ram Accelerators Paper 15 Univ ofWashington Seattle WA 1995

61Thaker A A and Chelliah H K ldquoNumerical Prediction of ObliqueDetonation Wave Structures Using Detailed and Reduced Reaction Mecha-nismsrdquo Combustion Theory and Modelling Vol 1 No 4 1997 pp 347ndash376

62Desbordes D Guerraud C and Hamada L ldquoSupersonic H2-AirCombustions Behind Oblique Shock Wavesrdquo 14th International Colloquiumon Dynamics of Explosions and Reactive Systems Paper D28 Univ ofCoimbra Portugal Aug 1993

63Li C Kailasanath K and Oran E S ldquoDetonation Structures Gener-ated by Multiple Shocks on Ram-Accelerator Projectilesrdquo Combustion andFlame Vol 108 1997 pp 173ndash186

64Viguier C and Desbordes D ldquoConditions of Onset and Stabilizationof Oblique Detonation Wavesrdquo Advances in Experimentation and Compu-tation of Detonations edited by G D Roy S H Frolov K Kailasanathand N Smirnov ENAS Moscow 1998 pp 40 41

65Bussing Tand Pappas G ldquoIntroduction toPulseDetonation EnginesrdquoAIAA Paper 94-0263 1994

66Helman D Shreeve R P and Eidelman S ldquoDetonation Pulse En-ginerdquo AIAA Paper 86-1683 June 1986

67Cambier J L and Adelman H G ldquoPreliminary Numerical Simu-lations of a Pulsed Detonation Wave Enginerdquo AIAA Paper 88-2960 July1988

68Eidelman S Grossmann W and Lottati I ldquoAir-Breathing PulsedDetonationEngine Concept A Numerical Studyrdquo AIAA Paper 90-2420 July1990

69Eidelman S Grossmann W and Lottati I ldquoComputational Analysisof Pulsed Detonation Engines and Applicationsrdquo AIAA Paper 90-0460 Jan1990

1702 KAILASANATH

Fig 10 Schematic of a ram accelerator in the detonative mode

Fig 11 Early concept of the detonative mode (from Ref 37)

a projectile that resembles the centerbody of a ramjet travels at asupersonic speed through a premixed fuelndashoxidizerndashdiluent mix-ture enclosed in a tube Because the projectile travels at supersonicspeeds oblique shock waves are formed at the nose of the projec-tile and they subsequently re ect from the side walls of the tubeand the projectile body Depending on the strength of these obliqueshocks ignition may occur after the initial shock or subsequentre ections Depending on the velocity of the projectile differentmodes of combustion and correspondingly different modes of op-eration of the system are possible When the velocity of the pro-jectile is greater than the CndashJ velocity an ODW may be stabilizedon the projectile resulting in the detonative mode of operation Aschematic of the ram accelerator operating in the detonative modeis shown in Fig 10 From the early days detonation has been dis-cussed as a means to attain very high velocities including escapevelocities37 In Ref 37 and several subsequent investigations thestructure of the detonative mode was described by the schematicdiagram shown in Fig 11 The projectile nose cone angle and thesound speed in the mixture selected were tailored to avoid combus-tion behind the initial shock wave but to provide initiation behindthe rst shock wave re ected from the tube wall Emphasis was alsoplaced on having the rst re ected shock wave strike the projec-tile body just aft of the shoulder37 If the system needed to be de-signed so precisely successful operation would indeed be dif cultin practical situations Further research has shown that the systemis far more stable and the particular scenario shown in Fig 11 ac-tually occurs just prior to system failure There were early reportsof successful attainment of superdetonative projectile velocities39

but the experimental emphasis shifted to the successful develop-ment of subdetonative modes of operation except at the Instituteof Saint-Louis (ISL) where research on the detonative mode hascontinued40 iexcl 42

Numerical simulations4344 and theoretical studies4546 based onsimpli ed analytical models of superdetonative ram acceleratorswere startedin the earlydays and have made asigni cant impactoverthe past decade These early studies con rmed the potential of ramaccelerators to accelerate ef ciently large masses to escape veloci-ties and low Earth orbit The steady-state simulations of Yungster47

highlighted the role of the projectile velocity (or Mach number)and showed that detonation could occur at different locations alongthe projectile body depending on the projectile Mach number (de- ned as the ratio of the velocity of the projectile to the speed ofsound in the premixed mixture) This work also raised the issue ofcombustion in the boundary layers along the projectile

The projectile velocity changes continuously in the ram acceler-ator and therefore steady-state simulations can at best be thoughtof as trying to characterize an instantaneous snapshot of the sys-tem The importance of the transient processes was rst highlightedin the work of Li et al48 This as well as subsequent work49 hasshown the importance of taking into account the time-dependent

Fig 12 Schematic showing the detailed structure of an oblique deto-nation and its relation to the detonation structures in ram accelerators

nature of the ow in the ram accelerator in describing phenomenasuch as unstarts and other system failures Computational studiesof the starting aspects of ram accelerators and the role of viscouseffects have also been conducted50 iexcl 52 Experimental efforts haveprimarily focused on the nondetonative mode except at ISL40 iexcl 42

Before discussing some of the experimental results and the dif cul-ties encountered in the practical development of ram accelerators inthe detonative mode two critical issues need to be discussed Oneissue is the actual structure of ODWs and the other is the stabilityof such detonation waves during transient operation

Structure and Stability of ODWsA key issue in the development of the ODWE and the ram accel-

erator is the structure and stability of ODWs Although the use ofODW was suggested as early as 195815 the stability of such waveswas implicitly assumed and questions about their detailed structurewere not considered Renewed interest in the late 1980s led to a the-oretical analysis of the supersonic ow of a combustible gas mixturepast a wedge53 Reference 53 indicated that for approach velocitiesroughly 25 or more larger than the CndashJ velocity of the reactantmixture ODWs can be stabilized over a useful range of wedge an-gles For wedge angles greater than a critical value the wave willdetach from the wedge and form an overdriven detonation wave ora more complex shock-induced combustion wave About this timethere were also numerical studies that showed that standing det-onation waves can be established on a wedge under certain owconditions5354

The early experimental efforts61314 were either focused on stabi-lizing a normal detonation wave in a duct or involved shock-inducedcombustion behind weak oblique shock waves Therefore it was notclear if ODWs could really be stabilized until some experiments inthe late 1980s on the diffraction and transmission of detonationsshowed that they could be5556

More recently the interest in ram accelerators has spawned alarge number of studies of the structure and stability of ODWs57 iexcl 61

The work of Li et al57 iexcl 59 showed that the basic structure of anoblique detonation generated by a wedge is more complex thanpreviously thought and consists of a nonreactive shock wave an in-duction region de agration waves and a detonation wave in whichthe pressure increase due to the energy release is closely coupledto the shock front This basic structure and a schematic of how itmight relate to the detonation waves in ram accelerator is shown inFig 12 These studies also showed that over a wide range of ow

KAILASANATH 1703

and mixture conditions the basic detonation structure is stable andvery resilient to disturbances in the ow In agreement with theory53

the simulations also showed the existence of a critical angle beyondwhich an attached stable detonation does not occur59 Other recentstudies have shown the same complex structure even with the inclu-sion of more complex chemical kinetic schemes6061 The essentialfeaturesof the complex ODW structure have also been con rmed byexperiments462 Thus a major conclusion from the recent researchis that the structure of an ODW can no longer be assumed to bea straight oblique shock wave followed closely by energy releaseand the complex structure can indeed be stabilizedon wedge-shapedobjects Implications of this work on propulsion applications mustbe considered

Although the schematic of Fig 12 shows how to relate the basicODW structure to that in ram accelerators the actual detonationwave structure in ram accelerators is likely to be more complexbecause of the expansion waves and multiple shock waves that aregenerated due to the geometric complexity of the system63 Theresults from a study of the detailed structure of detonations in ramaccelerators under a variety of ow conditions63 are summarizedin Fig 13 where the detonation wave structure is shown for three ow Mach numbers Recent experimental investigations have alsorevealed extremely complex detonation wave structures involvingmultiple shock waves under certain conditions64 In spite of theircomplexity all of these structures appear to be inherently relatedto the basic three-wave structure of oblique detonations shown inFig 12 However this appears to be an area of research that willcontinue for a while

a)

b)

c)

Fig 13 Detailed structure of detonations inram accelerators operatingat a) Mach 76 b) Mach 80 and c) Mach 84 (from Ref 63)

a) Case A Mach = 905

b) Case B Mach = 915

Fig 14 Visualization of the ow eld during the acceleration of theprojectile shown in Fig 5 from Mach 8 to 10 at two instants

Transient Nature of Detonation Waves in Ram AcceleratorsThe second important feature revealed by the recent studies is

the inherent transient nature of detonation waves in con gurationssuch as the ram accelerator When the concept of ram acceleratorsin the detonative mode was rst proposed it was thought that theleading shock wave (from the nose of the projectile) should re ectfrom the tube walls such that it impinges exactly at the shoulderof the projectile This will maximize the thrust and hence accel-eration However if the system were designed so precisely for onecondition its off-design performance could be unacceptable Fur-thermore simulations have shown that as the projectile acceleratesthe detonation wave structure tends to also move49 To illustrate thedynamics of detonations in the ram accelerator two photographsof the ow eld from a simulation (where the projectile acceleratedfrom Mach 8 to 10) are shown in Fig 14 In case A the projectileMach number is 905 and in case B it is 915 For both cases themain illustrations show the concentration of water vapor with thetwo inserts showing the pressure and temperature in a selected re-gion of the ow eld At the instant shown in Fig 14a (case A) thepressure and temperature behind the rst re ected shock wave arenot very high and combustion and energy release (as indicated bythe water vapor concentration) takes place only behind the secondre ected shock wave That is the detonation can be said to occurbehind the second re ected shock wave By the time the projectilehas accelerated to Mach 915 the leading shock wave has strength-ened raising the pressure and temperature behind the rst re ectedshock wave and causing the detonation to occur there

The behavior just discussedis seenthroughout the operationof theram accelerator in the detonative mode That is detonation waves inram accelerators are dynamic and move from shock wave to shockwave as the projectile acceleratesEventually the detonation wouldoccur just after the rst shock wave and the whole system will ceaseto operate because there will not be any positive thrust Furthermorethe detonation wave will run in front of the projectile creating a sit-uation similar to the classical unstart phenomena in ramjets Suchunstarts and detonations occurring after a different number of re- ections have also been noted in experiments42 Hence the shape ofthe projectile has to be chosen carefully so that the detonation canmove over the projectile body and still produce positive thrust overthe regime of interest A practical problem that has hindered thedevelopment of ram accelerators operating in the detonative modeis the change in the shape of the projectile due to material erosionburning and failure This is currently being overcome by the use ofsteel cowlings42

1704 KAILASANATH

PDEsThe other major current propulsion application involving detona-

tion waves is the PDE The basic concept behind a PDE is simpleDetonation is initiated repeatedly at either the closed or the openend of a detonation chamber that is lled with a premixed fuelndashairmixture Figure 15 (based on Ref 65) illustrates the concept forthe case when the detonation wave is initiated at the closed endAs shown in this schematic a one-dimensional planar detonation isinitiated near the closed end of the detonation chamber and travelsat the CndashJ velocity toward the open end A set of rarefaction wavestrail the detonation wave to reduce the velocity to zero at the closedend of the tube When the planar detonation wave exits the chamberanother set of expansion waves is generated that travels toward theclosed end These waves evacuate the burned detonation productsresulting in a cool empty chamber that is ready to be lled with afresh fuelndashair mixture The entire cycle repeats resulting in a pe-riodic high-pressure zone near the closed end of the chamber Theintegrated effectof this high pressure over the closed end (represent-ing a thrust wall) produces thrust This is of course a very idealizedpicture of the system Before discussing progress made in the re-cent years in developing a PDE a brief review of the background isprovided

The early work on the use of intermittent detonations has alreadybeen discussed In the late 1980s this concept of using intermittentor pulsed detonations was reexamined experimentally at the NavalPostgraduate School66 An ethylenendashoxygen detonation wave in asmall-diameter tube was used as a predetonator to initiate detona-tions in a larger tube containing an ethylenendashair mixture Periodicfuel injection within the naturally aspirated tube resulted in an in-termittent frequency of 25 Hz Speci c impulse estimated usingthe pressure time history and the amount of fuel consumed rangedfrom 1000 to 1400 s The velocities of the observed waves (lessthan 1 kms) are signi cantly below the CndashJ detonation wave ve-locities for the reported mixtures indicating that a fully developeddetonation wave was not formed

The rst purely computational study of the PDE reported in theopen literature appears to be the work of Cambier and Adelman67

The system simulated consisted of a 50-cm-long main tube attachedto a 43-cm-long diverging nozzle Quasi-one-dimensional simula-tions using the Euler equations for unsteady ow were carried outwith a total variation diminishing scheme and multistep nite ratekinetics A detonation was initiated at the closed end of a tube lledwith a premixed stoichiometric hydrogenndashair mixture at 3 atmOverall performance calculated by integrating the instantaneousthrust and the fuel ow rates gave a speci c impulse Isp of about6500 s and a range of operating frequencies up to 667 Hz Issues ofignition energies and transition to detonation were not consideredand the dif culties with attaining rapid mixing of fuel and air wasavoided by using a premixed fuelndashair mixture These assumptionsare still typical of those used in more recent simulations

The rst look at the interactions between the ow inside the cham-ber and that outside appears to be the work of Eidelman et al7

who simulated a cylindrical detonation chamber (about 15-cm diam

Fig 15 Schematic of an idealized PDE showing various stages duringone cycle of operation (adapted from Ref 65)

Fig 16 Generic PDE device used for somecomputational studieswheredetonation is initiated near the open end (from Ref 69)

and 15 cm long) with a small converging nozzle In their two-dimensional simulations a self-similarsolution for a planar detona-tion was imposed near the open end and made to travel toward theclosed end The re ection of the wave from the closed end createda higher pressure at the wall than during initiation near the closedend Hence this mode of initiation was considered to be better Af-ter the detonation left the chamber the pressure in the chamber wasobserved to fall below the ambient resulting in ingestion of outsideair In subsequent works86869 the same numerical techniques andmodels were used to investigate a slightly different con gurationwhere the nozzle at the aft end was removed and an inlet was in-troduced near the closed end This con guration has become oneof the popular systems for two-dimensional (and axisymmetric) nu-merical simulations Again detonation was assumed to be initiatedat the open end and traveled toward the closed end In this con gu-ration when the pressure falls below atmospheric in the detonationchamber air is ingested through the inlets in the front even duringstatic operations Thus a valveless self-aspiratingoperation is pos-sible These simulations also introduced the discussion of open-endinitiation vs closed-end initiation and their relative advantages Thecurrent status of this issue will be discussed later

Many of the early efforts on intermittent or PDEs have been re-viewed in Ref 7 and 8 where a link is also made between thesedetonation engines and pulse jet engines (which do not use detona-tive combustion) In addition Ref 8 provides a detailed summaryof the numerical investigations conducted by the authorsrsquo group ona generic PDE device shown in Fig 16 A characteristic feature ofthis device is that the detonation wave is initiated at the open endand travels toward a thrust wall at the closed end An advantage ofthis approach is that it is valveless and self-aspirating Air is en-trained into the chamber through inlets near the closed end whenthe pressure falls below atmospheric pressure during part of the cy-cle Unsteady Euler simulations were used to study the operationof the device from static (M = 0) to supersonic (M = 2) conditionsAs stated in Ref 8 a proof of principle experimental demonstra-tion of the PDE mode of operation where detonation is initiated atthe open end still needed to be done Furthermore issues such asdetonation initiation and structure and fuelndashair injection and mixingalso needed to be studied

Bussing and Pappas65 provided a detailed description of thebasic operation of an ideal zed PDE They also reported someone-dimensional studies of PDEs burning hydrogenndashoxygen andhydrogenndashair mixtures The detonation wave was initiated at theclosedend in this engine using a high-temperature and high-pressureregion An advantage of this design is the con nement provided atthe closed end which should enable a more rapid initiation andestablishment of a detonation Both airbreathing and rocket modeof operation of this valved design was discussed Lynch et al70

presented a computational study of an axisymmetric PDE with astraight inlet very similar to that of Eidelman and Grossmann8 andEidelman et al69 Various inlet lengths from 1 to 12 cm and ightMach numbers of 08 and 20 were studied The results were similarto those reported earlier except that some additional details of thedetonation wave front and the air induction system were resolved inthese studies In a related work Lynch and Edelman71 showed thatreductions in scavenging time (and hence overall cycle time) canbe achieved by suitably shaping the inlet

Bussing et al72 compared open-end initiation and closed-end ini-tiation in a simple tube PDE geometry Equivalent thrust productionand fuel ef ciency were observed for the two types of initiation

KAILASANATH 1705

Although a higher head-end pressure was attained during open-endinitiation the pressure continuously dropped for that case unlike theclosed-end initiation where a constant pressure region occurred fora while before reduction of the pressure to the ambient value Morerecently an experimental study has con rmed the differences in theobserved pressure pro les and it has also been concluded that themaximum level of impulse attained is independent of the locationof the detonation initiator73

In the past few years there has been a substantial growth in theappearance of papers related to the PDE and complete sessionshave been devoted to the topic at annual AIAAASMESAEASEEjoint propulsion conferences Because of space considerations onlyselected papers from these years are brie y discussed here Sterlinget al74 conducted one-dimensional numerical investigations of aself-aspiratingPDE operating for multiple cyclesA key observationfrom their studies was that only a portion of the detonation tubecan be lled with fresh charge under self-aspirating operation Inspite of this limitation they reported a speci c impulse of 5151 sfor a hydrogenndashair system However they concluded that the idealperformance of such an engine is ldquonear those of other hydrogen-fueledair breathing enginesrdquo74

Several basic shock tube experiments related to the developmentof a hydrogen-fueled PDE were described by Hinkey et al75 Theseinclude the measurement of detonation wave velocities and de a-gration to detonation transition (DDT) lengths for a range of equiv-alence ratios As expected the DDT lengths were found to be toolarge for practical applications even in hydrogenndashoxygen mixturesTherefore traditional techniques such as the use of a Schelkin spiral(see Ref 76) were tried to reduce the transition length A factor of2ndash4 reduction was observed over the range of equivalence ratios in-vestigated This work75 highlights the dif culties involved in obtain-ing a fully developed detonation in a short distance (as assumed inthe conceptual studies and numerical simulations discussed earlier)and questions if the detonations reported in previous experimentalinvestigations161766 were really fully transitioned from de agra-tions The speci c impulses measured in their experiments (about240 s for hydrogenndashoxygen and 1200 s for hydrogenndashair) were alsostated to be in agreement with their analysis

An interesting PDE concept presented by Bussing77 was that of arotary-valved multiple-pulsed detonation engine Here several det-onation chambers are coupled to an air inlet and fuel source usinga rotary valve as suggested in one of the early papers of Nichollset al16 The rotary valves allow the lling of some detonation cham-bers while others are detonating or exhausting Other papers includepresentation of an overall PDE performance model78 and the exper-imental characterization of the detonation properties of some fuelsfor use in PDEs79

The importance of adequately mixing the fuel and oxidizer washighlighted by the experimental investigations of Stanley et al80

who obtained very low sub-CndashJ velocities when injecting the fueland oxidizer at different times and not taking extra effort to en-sure that they were well mixed The use of turbulence producingdevices appeared to improve signi cantly the mixing and the at-tainment of higher velocities but also resulted in signi cant thrustlosses This study also showed that raising the initial pressure wasbene cial The computational studies of Eidelman et al81 showedthat the PDE engine could operate even for a range of transitionaldetonation regimes that produced nonplanar or not fully developeddetonations Aarnio et al82 discussed two failure modes DDT tran-sition failure and premature ignition In both cases the system con-tinued to operate though there would be loss of thrust In Ref 82detailed discussions of the pressure and thrust histories during a5-Hz 20-cycle operation are also provided (Fig 17) The cycle tocycle variations of peak pressure have been attributed to the lowsampling frequency used82 Both a direct measurement as well asthe integration of the pressure histories were used to estimate the av-erage Isp from hydrogenndashair systems to be between 1116 and 1333 sThe effect of the predetonator volume on the reduction of speci cimpulse was also shown to be a critical issue The effects of inletsand nozzles were not considered in this study

A conceptual design of a multitube PDE with a single air inletduct was discussed by Pegg et al83 Time-dependent computational

Fig 17 Pressure history at one location from a PDE operating at 5 Hzfor 20 cycles (from Ref 82)

uid dynamics analysis indicated that the inlet isolaterdiffusercon-cept would work and did not allow enough time for the formationof destabilizing hammer shocks Performance analysis includingcomponent ef ciencies reported in the paper indicate that opera-tional frequencies of the order of 75ndash100 Hz are required for theMach 12ndash3 ights considered Experimental data indicating mul-ticycle operation at 100 Hz with a rich ethylenendashoxygen mixturewere presented by Sterling et al84 DDT enhancement devices andpredetonators were used in the experiments In spite of this det-onation waves did not occur all of the time and when detonationwaves failed the tube became very hot and testing was suspendedTherefore the maximum testing time was reported to be 05 s Inspite of this limitation the two studies8384 taken together high-light the possibility of PDEs burning hydrocarbon fuels for realisticmissions

Five different nozzle shapes and their effect on performance werestudied computationally by Cambier and Tegner85 Their results in-dicate that the presence of a nozzle can affect the performance ofthe PDE by increasing the thrust delivery during the ignition phaseThe bell-shaped nozzles appeared to give higher performance thanshapes with positive curvature Their results also showed that noz-zles also affected the ow dynamics and hence the timing of thevarious phases of the engine cycle In an experimental study of aconical exhaust nozzle the authors of Ref 86 noted improved per-formance (based on higher velocities which were still sub-CndashJ) butdid not see any effect on the blowdown process More recently87

the effectof various nozzle shapes including converging divergingand straight have been reexamined computationally The converg-ing sections of nozzles introduced shock wave re ections whereasdiverging sections generated a negative thrust for a portion of thecycle due to overexpansion In spite of these limitations the overallconclusion of these studies was that ldquonozzles can drastically in-crease ef ciency of the PDEsrdquo87 There is also some experimentalevidence suggesting that propulsive performances can be increasedby the addition of a nozzle73 However factors such as the effect ofthe nozzles on detonation transmission and the detailed dynamicsof the ow have not yet been elucidated This appears to be an areathat needs further investigation

Other recent efforts have focused on demonstrating the opera-tion of single- and multitube combustors coupled to an inlet using arotary-valvemechanism88 The valve servesto both meter the air owinto the combustor and to isolate the inlet from the high pressuresproduced during the detonation cycle Utilizing multiple combus-tors that ll and detonate out of phase allows the continual use of thein owing air Firing rates of up to 12 Hz per combustor were demon-strated for a hydrogen-fueled system Several other issues related tothe practical development of PDEs have been discussed by Bussinget al89 According to them three of the more important engineer-ing issues are the development of a valve-injection subsystem anignition system and a low-volume predetonator

The use of PDEs for rockets has been revived recently90 Therocket mode of operation is very similar to the description providedearlier of the airbreathing engine with ignition at the closed endexcept that the oxidizer also needs to be injected into the system pe-riodically An advantage of PDEs for rockets would be their higherpower density thus enabling the development of more compact

1706 KAILASANATH

rockets Reference 90 shows the pressure traces from a pulsed det-onation rocket engine operating at 145 Hz on a hydrogenndashoxygenmixture

More recently a two-tube rotary-valved PDE with ight-sizecomponents was operated at 40 Hz per combustor for 30 s withan ethylenendashair mixture91 In addition throttling of the device al-lowed operation at three thrust levels during a 10-s demonstrationexperiment Contrary to some earlier estimates the thermal loadswere signi cant and it was not possible to operate for more than3 s without water cooling It is not clear if this was because ofperiodic detonation failures and consequent de agrative modes ofcombustion as reported in some earlier studies

The work reported in the open literature and discussed earlier hasfocused on gaseous fuels However for volume-limited propulsionapplications PDEs operating on liquid fuels need to be demon-strated In a recent study Brophy et al92 reported on experimentsusing JP-10oxygen and JP-10air aerosols The fuelndashoxygen mix-ture was successfully detonated to obtain an engine operating at5 Hz but the tests with the fuelndashair mixture were not successful

Computational studies continue to make progress in the study ofthe PDE8793 iexcl 97 The effect of nozzle shapes and incomplete transi-tion to CndashJ detonations have been studied by Eidelman and Yang87

They nd that the cycle ef ciency will be virtually the same whethera CndashJ detonation wave is initiated instantaneously or if the CndashJ valueis attainedonly at the end of the detonation chamber Primarilybasedon this study87 they come to an interesting conclusion that ldquoinitia-tion energies that are required for PDE operations can be small andcomparable to energies required for initiation of (other) combustionprocessesrdquo However this is yet to be substantiated experimentallySekar et al93 reexamine open- and closed-end initiation and thebasic self-aspirating con guration of Eidelman8 using more recentcomputational tools Injection and mixing issues that have gener-ally been ignored in previous numerical studies are also beginningto be addressed94 Both front-wall and lateral-wall injections werestudied and the time required for fuel injection was estimated Itis suggested that it may be possible to ll only a portion of thechamber during high-frequency operations Recent computationalstudies also suggest that preconditioning the fuelndashair mixture us-ing a shock wave can signi cantly reduce the DDT distance95 Theperformance estimates mentioned throughout the discussion of thePDE show signi cant variation A systematic review of various per-formance estimates and possible reasons for the observed variationhave been presented9697 These numerical studies suggest that thetime history of the back pressure is a crucial parameter affecting theestimated performance

Other Detonation EnginesThere have also been several other interesting conceptual studies

on the applications of detonations to engines Cambier et al98 haveproposed a pulsed-detonation wave augmentation device as part ofa hybrid engine for a single-stage-to-orbit airbreathing hypersonicvehicle Here the PDE is used for both thrust generation as well asfor mixingcombustion augmentation Two speci c con gurationswere investigated numerically

The concept of a detonation internal combustion engine was pro-posed by Loth and Loth99 and Loth100 Here a conventional pistoncylinder con guration is modi ed to incorporate a separate deto-nation chamber which is isolated by a valve from the compressionchamber during fuel injection and discharges tangentially into a vor-texchamber formedby the piston and cylinder at top deadcenterTherapid detonative combustion is followed by rapid dilution throughmixing with air in the vortex chamber to reduce the formation ofNOx and unburned hydrocarbons and to achieve an overall leanmixture The vortex chamber was also designed to store a portion ofthe detonation waversquos kinetic energy Overall thermal ef cienciessomewhat better than constant volume combustion were postulatedunder ideal conditions

ConclusionsDetonation waves have been explored extensively for propulsion

applications because of their inherent theoretical advantage overde agrative combustion However practical developments to date

have been of idealized systems or laboratory-scale devices or bothThe basic ideas behind most of the current systems being investi-gated have been known for a while but still many of the details needto be understood Advances in computations and experimental di-agnostics appear to be poised to make practical propulsion devicesbased on detonation waves a reality

Two major systems currently under development are the ram ac-celerator and the PDE In both cases the theoretical and computa-tional studies have been far more encouraging than the experimentsHowever in both cases actual demonstrations of devices workingon gaseous fuels have taken place In the case of the ram accelera-tor fuelndashoxygenndashdiluent mixtures have been used and the primaryproblem appears to be heat transfer and related material failuresThe use of gaseous fuels is not a major issue here but new materialsfor the projectiles need to be explored

In the case of the PDE most of the successful operations havebeen with fuelndashoxygen mixtures There still has not been a directcomparison between detailed experimental data and related theoret-ical and computational results Furthermore most of the PDE andrelated work to date have focused on gaseous mixtures PDEs op-erating on multiphase mixtures need to be emphasized because itwould be more practical to use liquid fuels for most of the propul-sion applicationsproposed for these systemsThere are several othergeneral issues that need to be resolved Clearly the evacuation ofexhaust gases is important both from the point of view of controllingthe cycle time and for preventing autoignition and mixing with freshgases The heat transfer and noise from PDEs have also not beenreported in any detail Other issues include inletndashcombustor andcombustorndashnozzle interactions and scalability of the entire device

AcknowledgmentsThis work has been sponsored by the Of ce of Naval Research

through the Mechanics and Energy Conversion Division and theUS Naval Research Laboratory

References1Fickett W and Davis W C Detonation Univ of California Press

Berkeley CA 19792Hoffmann N ldquoReaction Propulsion by Intermittent Detonative Com-

bustionrdquo German Ministry of Supply AI152365 Volkenrode Translation1940 (cited in Ref 16)

3Roy M ldquoPropulsion par Statoreacteur a Detonationrdquo Comptes-Rendusde lrsquoAcademie des Sciences Vol 222 1946 pp 31 32

4Dabora E K and Broda J C ldquoStanding Normal Detonations andOblique Detonations for Propulsionrdquo AIAA Paper 93-2325 July 1993

5Dabora E K ldquoStatus of Gaseous Detonation Waves and Their Rolein Propulsionrdquo Fall Technical Meeting of the Eastern States Section ofthe Combustion Institute Combustion Inst Pittsburgh PA 1994 pp 11ndash

186Rubins P M and Bauer R C ldquoReview of Shock-Induced Supersonic

Combustion Research and Hypersonic Applicationsrdquo Journal of Propulsionand Power Vol 10 No 5 1994 pp 593ndash601

7Eidelman S Grossmann W and Lottati I ldquoReview of PropulsionApplications and Numerical Simulations of the Pulsed Detonation EngineConceptrdquo Journal of Propulsion and Power Vol 7 No 6 1991 pp 857ndash

865 also AIAA Paper 89-2446 July 19898Eidelman S and Grossmann W ldquoPulsed Detonation Engine Experi-

mental and Theoretical Reviewrdquo AIAA Paper 92-3168 July 19929Reingold L ldquoRecherches sur les combustions permanentes apportees

aux foyers a circulation interne supersoniquerdquo ONERA TN2 Dept of En-ergy and Propulsion Study 728-E Paris March 1950

10Bitondo D and Bollay W ldquoPreliminary Performance Analysis ofthe Pulse-Detonation-Jet Engine Systemrdquo Aerophysics Development CorpRept ADC-102-1 April 1952

11Gross R A ldquoExploratory Study of Combustion in Supersonic FlowrdquoUS Air Force Of ce of Scienti c Research TN 59-587 Washington DCJune 1959

12Gross R A and Chinitz W ldquoA Study of Supersonic CombustionrdquoJournal of Aerospace Science Vol 27 No 7 1960 pp 517ndash525

13Nicholls J A Dabora E K and Gealer R L ldquoStudies in Con-nection with Stabilized Gaseous Detonation Wavesrdquo Seventh Symposium(International) on Combustion Combustion Inst Pittsburgh PA 1959 pp766ndash772

14Nicholls J A and Dabora E K ldquoRecent Results on Standing Detona-tion Wavesrdquo Eighth Symposium (International) on Combustion CombustionInst Pittsburgh PA 1962 pp 644ndash655

KAILASANATH 1707

15Dunlap R Brehm R L and Nicholls J A ldquoA Preliminary Studyof the Application of Steady-State Detonative Combustion to a ReactionEnginerdquo Jet Propulsion Vol 28 No 7 1958 pp 451ndash456

16Nicholls J A Wilkinson H R and Morrison R B ldquoIntermittentDetonation as a Thrust-Producing Mechanismrdquo Jet Propulsion Vol 27No 5 1957 pp 534ndash541

17Krzycki L J ldquoPerformance Characteristics of an Intermittent Detona-tion Devicerdquo US Naval Ordnance Test Station US Naval Weapons Rept7655 China Lake CA 1962

18Nicholls J A Cullen R E and Ragland K W ldquoFeasibility Studiesof a Rotating Detonation Wave Rocket Motorrdquo Journal of Spacecraft andRockets Vol 3 No 6 1966 pp 893ndash898

19Adamson T C and Olsson G R ldquoPerformance Analysis of a RotatingDetonation Wave Rocket Enginerdquo Astronautica Acta Vol 13 No 4 1967pp 405ndash415

20Shen P I and Adamson T C ldquoTheoretical Analysis of a RotatingTwo-Phase Detonation in Liquid Rocket Motorsrdquo Astronautica Acta Vol17 No 4 1972 pp 715ndash728

21Clayton R M and Rogero R S ldquoExperimental Measurements on aRotating Detonation-Like Wave Observed During Liquid Rocket ResonantCombustionrdquo Jet Propulsion Lab TR 32-788 California Inst of Technol-ogy Pasadena CA Aug 1965

22Back L H ldquoApplication of Blast Wave Theory to Explosive Propul-sionrdquo Acta Astronautica Vol 2 1975 pp 391ndash407

23Varsi G Back L H and Kim K ldquoBlast Wave in a Nozzle for Propul-sion Applicationsrdquo Acta Astronautica Vol 3 1976 pp 141ndash156

24Kim K Varsi G and Back LH ldquoBlast Wave Analysis for DetonationPropulsionrdquo AIAA Journal Vol 10 No 10 1977 pp 1500ndash1502

25Back L H Dowler W L and Varsi G ldquoDetonation PropulsionExperiments and Theoryrdquo AIAA Journal Vol 21 No 10 1983 pp 1418ndash

142726Kantrowitz A ldquoPropulsion to Orbit by Ground Based Lasersrdquo Astro-

nautics and Aeronautics Vol 10 No 5 1972 p 7427Hyde R A ldquoOne-D Modelling of a Two-Pulse LSD Thrusterrdquo Pro-

ceedings of the 1986 SDIODARPA Workshop on Laser Propulsion editedby J T Kare Lawrence Livermore National Labs 1987 pp 79ndash88

28Emery M H Kailasanath K Oran E S and Gardner J H ldquoCompu-tational Studies of Laser Supported Detonationsrdquo Proceedings of the 1987SDIO Workshop on Laser Propulsion edited by J T Kare Lawrence Liv-ermore National Labs 1990 pp 43ndash55

29Emery M H Kailasanath K and Oran E S ldquoNumerical Simu-lations of Laser Supported Detonationsrdquo 1988 SDIO Workshop on LaserPropulsion Lawrence Livermore National Labs April 1988

30Carrier G F Fendell F McGregor D Cook S and Vazirani MldquoLaser-Initiated Conical Detonation Wave for Supersonic CombustionrdquoAIAA Paper 91-0578 Jan 1991

31Fendell F E Mitchell J McGregor R Magiawala K and Shef eldM ldquoLaser-Initiated Conical Detonation Wave for Supersonic CombustionIIrdquo AIAA Paper 92-0088 Jan 1992

32Carrier G F Fendell F E and Chou M S ldquoLaser-Initiated ConicalDetonation Wave for Supersonic Combustion IIIrdquo AIAA Paper 92-3247July 1992

33Menees G P Adelman H G Cambier J-L and Bowles J V ldquoWaveCombustors for Trans-Atmospheric Vehiclesrdquo Journal of Propulsion andPower Vol 8 No 3 1992 pp 709ndash713

34Cambier J L Adelman H and Menees G ldquoNumerical Simulationsof an Oblique Detonation Wave Enginerdquo AIAA Paper 88-0063 1988

35Atamanchuk T and Sislian J ldquoOn- and Off-design Performance Anal-ysis of Hypersonic Detonation Wave Ramjetsrdquo AIAA Paper 90-2473 1990

36Kuznetsov M M Neyland V J and Sayapin G N ldquoEf ciency In-vestigation of Scramjet with Detonation and Shockless Supersonic Combus-tionrdquo Utchenye Zapisky TsAGI Vol 23 No 2 1992 pp 30ndash37

37Bruckner A P and Hertzberg A ldquoRam Accelerator Direct LaunchSystem for Space Cargordquo International Astronautical Federation Paper 87-211 1987

38Hertzberg A Bruckner A P and Bogdanoff D W ldquoRam Accel-erator A New Chemical Method for Accelerating Projectiles to UltrahighVelocitiesrdquo AIAA Journal Vol 26 No 2 1988 pp 195ndash203

39Kull A Burnham E Knowlen C Bruckner A P and HertzbergA ldquoExperimental Studies of the Superdetonative Ram Accelerator ModesrdquoAIAA Paper 89-2632 1989

40Patz G Seiler F Smeets G and Srulijes J ldquoStatus of ISLrsquos RA-MAC30 with Fin Guided Projectiles Accelerated in a Smooth Borerdquo Pro-ceedings of the Second International Workshop on Ram Accelerators Univof Washington Seattle WA 1995

41Seiler F Patz G Smeets G and Srulijes J ldquoGasdynamic Limitsof Ignition and Combustion of a Gas Mixture in ISLrsquos RAMAC30 ScramAcceleratorrdquo Proceedings of the Second International Workshop on RamAccelerators Univ of Washington Seattle WA 1995

42Seiler F Patz G Smeets G and Srulijes J ldquoIn uence of ProjectileMaterial and Gas Composition on Superdetonative Combustion in ISLrsquosRAMAC30rdquo AIAA Paper 98-3445 July 1998

43Brackett D C and Bogdanoff D W ldquoComputational Investigation ofOblique Detonation Ramjet-in-Tube Conceptsrdquo Journal of Propulsion andPower Vol 5 No 3 1989 pp 276ndash281

44Yungster S Eberhardt S and Bruckner A P ldquoNumerical Simula-tion of Shock-Induced Combustion Generated by High-Speed Projectiles inDetonable Gas Mixturesrdquo AIAA Paper 89-0673 1989

45Rom J and Kivity Y ldquoAccelerating Projectiles Up to 12 kms Utiliz-ing the Continuous Detonation Propulsion Methodrdquo AIAA Paper 88-29691988

46Humphrey J W ldquoParametric Study of an ODW Scramaccelerator forHypersonic Test Facilitiesrdquo AIAA Paper 90-2470 1990

47Yungster S ldquoNavierndashStokes Simulation of the Supersonic CombustionFlow eld in a Ram Acceleratorrdquo AIAA Paper 91-1916 1991

48Li C Kailasanath K Oran E S Boris J P Paper and LandsbergA M ldquoNumerical Simulations of Transient Flows in Ram AcceleratorsrdquoAIAA Paper 93-1916 1993

49Li C Kailasanath K Oran E S Landsberg A M and Boris J PldquoDynamics of Oblique Detonations in Ram Acceleratorsrdquo Shock Waves Vol5 June 1995 pp 97ndash101

50Yungster S and Radhakrishnan K ldquoComputational Study of FlowEstablishment in a Ram Acceleratorrdquo AIAA Paper 98-2489 July 1995

51Choi J Y Jeung I-S and Yoon Y ldquoNumerical Study of ScramAccelerator Starting Characteristicsrdquo AIAA Journal Vol 36 No 6 1998pp 1029ndash1038

52Bezgin L Ganzhelo A Gouskov O and Kopchenov V ldquoSome Nu-merical Investigation Results of Shock-Induced Combustionrdquo Proceedingsof the AIAA 8th International Space Planes and Hypersonic Systems andTechnologies Conference AIAA Reston VA 1998 pp 72ndash82

53Pratt D W Humphrey J W and Glenn D E ldquoMorphology of Stand-ing Oblique Detonation Wavesrdquo Journal of Propulsion and Power Vol 7No 5 1991

54Fujiwara T Matsuo A and Nomoto H ldquoTwo-Dimensional Detona-tion Supported by a Blunt Body or a Wedgerdquo AIAA Paper 88-0089 Jan1988

55Fan B C Sichel M and Kauffman C W ldquoAnalysis of ObliqueShock-Detonation Wave Interactions in the Supersonic Flow of a Com-bustible Mediumrdquo AIAA Paper 88-0441 Jan 1988

56Dabora E K Desbordes D Guerraud C and Wagner H GldquoOblique Detonations at Hypersonic Velocitiesrdquo Dynamics of Detonationsand Explosions Detonations Vol 133 Progress in Astronautics and Aero-nautics AIAA Washington DC 1990 pp 187ndash201

57Li C Kailasanath K and Oran E S ldquoStability of Oblique Detona-tions in Ram Acceleratorsrdquo AIAA Paper 92-0089 Jan 1992

58Li C Kailasanath K and Oran E S ldquoStructure of Reaction WavesBehind Oblique Shocksrdquo Dynamic Aspects of Detonation Vol 153 Progressin Astronautics and Aeronautics AIAA Washington DC 1993 pp 231ndash

24059Li C Kailasanath K and Oran E S ldquoDetonation Structures Behind

Oblique Shocksrdquo Physics of Fluids Vol 6 April 1994 pp 1600ndash161160Da Silva L F and Deshaies B ldquoNumerical Study of Some Ignition

Regimes of Combustible Supersonic Flows over a Wedgerdquo Proceedings ofSecond International Workshop on Ram Accelerators Paper 15 Univ ofWashington Seattle WA 1995

61Thaker A A and Chelliah H K ldquoNumerical Prediction of ObliqueDetonation Wave Structures Using Detailed and Reduced Reaction Mecha-nismsrdquo Combustion Theory and Modelling Vol 1 No 4 1997 pp 347ndash376

62Desbordes D Guerraud C and Hamada L ldquoSupersonic H2-AirCombustions Behind Oblique Shock Wavesrdquo 14th International Colloquiumon Dynamics of Explosions and Reactive Systems Paper D28 Univ ofCoimbra Portugal Aug 1993

63Li C Kailasanath K and Oran E S ldquoDetonation Structures Gener-ated by Multiple Shocks on Ram-Accelerator Projectilesrdquo Combustion andFlame Vol 108 1997 pp 173ndash186

64Viguier C and Desbordes D ldquoConditions of Onset and Stabilizationof Oblique Detonation Wavesrdquo Advances in Experimentation and Compu-tation of Detonations edited by G D Roy S H Frolov K Kailasanathand N Smirnov ENAS Moscow 1998 pp 40 41

65Bussing Tand Pappas G ldquoIntroduction toPulseDetonation EnginesrdquoAIAA Paper 94-0263 1994

66Helman D Shreeve R P and Eidelman S ldquoDetonation Pulse En-ginerdquo AIAA Paper 86-1683 June 1986

67Cambier J L and Adelman H G ldquoPreliminary Numerical Simu-lations of a Pulsed Detonation Wave Enginerdquo AIAA Paper 88-2960 July1988

68Eidelman S Grossmann W and Lottati I ldquoAir-Breathing PulsedDetonationEngine Concept A Numerical Studyrdquo AIAA Paper 90-2420 July1990

69Eidelman S Grossmann W and Lottati I ldquoComputational Analysisof Pulsed Detonation Engines and Applicationsrdquo AIAA Paper 90-0460 Jan1990

KAILASANATH 1703

and mixture conditions the basic detonation structure is stable andvery resilient to disturbances in the ow In agreement with theory53

the simulations also showed the existence of a critical angle beyondwhich an attached stable detonation does not occur59 Other recentstudies have shown the same complex structure even with the inclu-sion of more complex chemical kinetic schemes6061 The essentialfeaturesof the complex ODW structure have also been con rmed byexperiments462 Thus a major conclusion from the recent researchis that the structure of an ODW can no longer be assumed to bea straight oblique shock wave followed closely by energy releaseand the complex structure can indeed be stabilizedon wedge-shapedobjects Implications of this work on propulsion applications mustbe considered

Although the schematic of Fig 12 shows how to relate the basicODW structure to that in ram accelerators the actual detonationwave structure in ram accelerators is likely to be more complexbecause of the expansion waves and multiple shock waves that aregenerated due to the geometric complexity of the system63 Theresults from a study of the detailed structure of detonations in ramaccelerators under a variety of ow conditions63 are summarizedin Fig 13 where the detonation wave structure is shown for three ow Mach numbers Recent experimental investigations have alsorevealed extremely complex detonation wave structures involvingmultiple shock waves under certain conditions64 In spite of theircomplexity all of these structures appear to be inherently relatedto the basic three-wave structure of oblique detonations shown inFig 12 However this appears to be an area of research that willcontinue for a while

a)

b)

c)

Fig 13 Detailed structure of detonations inram accelerators operatingat a) Mach 76 b) Mach 80 and c) Mach 84 (from Ref 63)

a) Case A Mach = 905

b) Case B Mach = 915

Fig 14 Visualization of the ow eld during the acceleration of theprojectile shown in Fig 5 from Mach 8 to 10 at two instants

Transient Nature of Detonation Waves in Ram AcceleratorsThe second important feature revealed by the recent studies is

the inherent transient nature of detonation waves in con gurationssuch as the ram accelerator When the concept of ram acceleratorsin the detonative mode was rst proposed it was thought that theleading shock wave (from the nose of the projectile) should re ectfrom the tube walls such that it impinges exactly at the shoulderof the projectile This will maximize the thrust and hence accel-eration However if the system were designed so precisely for onecondition its off-design performance could be unacceptable Fur-thermore simulations have shown that as the projectile acceleratesthe detonation wave structure tends to also move49 To illustrate thedynamics of detonations in the ram accelerator two photographsof the ow eld from a simulation (where the projectile acceleratedfrom Mach 8 to 10) are shown in Fig 14 In case A the projectileMach number is 905 and in case B it is 915 For both cases themain illustrations show the concentration of water vapor with thetwo inserts showing the pressure and temperature in a selected re-gion of the ow eld At the instant shown in Fig 14a (case A) thepressure and temperature behind the rst re ected shock wave arenot very high and combustion and energy release (as indicated bythe water vapor concentration) takes place only behind the secondre ected shock wave That is the detonation can be said to occurbehind the second re ected shock wave By the time the projectilehas accelerated to Mach 915 the leading shock wave has strength-ened raising the pressure and temperature behind the rst re ectedshock wave and causing the detonation to occur there

The behavior just discussedis seenthroughout the operationof theram accelerator in the detonative mode That is detonation waves inram accelerators are dynamic and move from shock wave to shockwave as the projectile acceleratesEventually the detonation wouldoccur just after the rst shock wave and the whole system will ceaseto operate because there will not be any positive thrust Furthermorethe detonation wave will run in front of the projectile creating a sit-uation similar to the classical unstart phenomena in ramjets Suchunstarts and detonations occurring after a different number of re- ections have also been noted in experiments42 Hence the shape ofthe projectile has to be chosen carefully so that the detonation canmove over the projectile body and still produce positive thrust overthe regime of interest A practical problem that has hindered thedevelopment of ram accelerators operating in the detonative modeis the change in the shape of the projectile due to material erosionburning and failure This is currently being overcome by the use ofsteel cowlings42

1704 KAILASANATH

PDEsThe other major current propulsion application involving detona-

tion waves is the PDE The basic concept behind a PDE is simpleDetonation is initiated repeatedly at either the closed or the openend of a detonation chamber that is lled with a premixed fuelndashairmixture Figure 15 (based on Ref 65) illustrates the concept forthe case when the detonation wave is initiated at the closed endAs shown in this schematic a one-dimensional planar detonation isinitiated near the closed end of the detonation chamber and travelsat the CndashJ velocity toward the open end A set of rarefaction wavestrail the detonation wave to reduce the velocity to zero at the closedend of the tube When the planar detonation wave exits the chamberanother set of expansion waves is generated that travels toward theclosed end These waves evacuate the burned detonation productsresulting in a cool empty chamber that is ready to be lled with afresh fuelndashair mixture The entire cycle repeats resulting in a pe-riodic high-pressure zone near the closed end of the chamber Theintegrated effectof this high pressure over the closed end (represent-ing a thrust wall) produces thrust This is of course a very idealizedpicture of the system Before discussing progress made in the re-cent years in developing a PDE a brief review of the background isprovided

The early work on the use of intermittent detonations has alreadybeen discussed In the late 1980s this concept of using intermittentor pulsed detonations was reexamined experimentally at the NavalPostgraduate School66 An ethylenendashoxygen detonation wave in asmall-diameter tube was used as a predetonator to initiate detona-tions in a larger tube containing an ethylenendashair mixture Periodicfuel injection within the naturally aspirated tube resulted in an in-termittent frequency of 25 Hz Speci c impulse estimated usingthe pressure time history and the amount of fuel consumed rangedfrom 1000 to 1400 s The velocities of the observed waves (lessthan 1 kms) are signi cantly below the CndashJ detonation wave ve-locities for the reported mixtures indicating that a fully developeddetonation wave was not formed

The rst purely computational study of the PDE reported in theopen literature appears to be the work of Cambier and Adelman67

The system simulated consisted of a 50-cm-long main tube attachedto a 43-cm-long diverging nozzle Quasi-one-dimensional simula-tions using the Euler equations for unsteady ow were carried outwith a total variation diminishing scheme and multistep nite ratekinetics A detonation was initiated at the closed end of a tube lledwith a premixed stoichiometric hydrogenndashair mixture at 3 atmOverall performance calculated by integrating the instantaneousthrust and the fuel ow rates gave a speci c impulse Isp of about6500 s and a range of operating frequencies up to 667 Hz Issues ofignition energies and transition to detonation were not consideredand the dif culties with attaining rapid mixing of fuel and air wasavoided by using a premixed fuelndashair mixture These assumptionsare still typical of those used in more recent simulations

The rst look at the interactions between the ow inside the cham-ber and that outside appears to be the work of Eidelman et al7

who simulated a cylindrical detonation chamber (about 15-cm diam

Fig 15 Schematic of an idealized PDE showing various stages duringone cycle of operation (adapted from Ref 65)

Fig 16 Generic PDE device used for somecomputational studieswheredetonation is initiated near the open end (from Ref 69)

and 15 cm long) with a small converging nozzle In their two-dimensional simulations a self-similarsolution for a planar detona-tion was imposed near the open end and made to travel toward theclosed end The re ection of the wave from the closed end createda higher pressure at the wall than during initiation near the closedend Hence this mode of initiation was considered to be better Af-ter the detonation left the chamber the pressure in the chamber wasobserved to fall below the ambient resulting in ingestion of outsideair In subsequent works86869 the same numerical techniques andmodels were used to investigate a slightly different con gurationwhere the nozzle at the aft end was removed and an inlet was in-troduced near the closed end This con guration has become oneof the popular systems for two-dimensional (and axisymmetric) nu-merical simulations Again detonation was assumed to be initiatedat the open end and traveled toward the closed end In this con gu-ration when the pressure falls below atmospheric in the detonationchamber air is ingested through the inlets in the front even duringstatic operations Thus a valveless self-aspiratingoperation is pos-sible These simulations also introduced the discussion of open-endinitiation vs closed-end initiation and their relative advantages Thecurrent status of this issue will be discussed later

Many of the early efforts on intermittent or PDEs have been re-viewed in Ref 7 and 8 where a link is also made between thesedetonation engines and pulse jet engines (which do not use detona-tive combustion) In addition Ref 8 provides a detailed summaryof the numerical investigations conducted by the authorsrsquo group ona generic PDE device shown in Fig 16 A characteristic feature ofthis device is that the detonation wave is initiated at the open endand travels toward a thrust wall at the closed end An advantage ofthis approach is that it is valveless and self-aspirating Air is en-trained into the chamber through inlets near the closed end whenthe pressure falls below atmospheric pressure during part of the cy-cle Unsteady Euler simulations were used to study the operationof the device from static (M = 0) to supersonic (M = 2) conditionsAs stated in Ref 8 a proof of principle experimental demonstra-tion of the PDE mode of operation where detonation is initiated atthe open end still needed to be done Furthermore issues such asdetonation initiation and structure and fuelndashair injection and mixingalso needed to be studied

Bussing and Pappas65 provided a detailed description of thebasic operation of an ideal zed PDE They also reported someone-dimensional studies of PDEs burning hydrogenndashoxygen andhydrogenndashair mixtures The detonation wave was initiated at theclosedend in this engine using a high-temperature and high-pressureregion An advantage of this design is the con nement provided atthe closed end which should enable a more rapid initiation andestablishment of a detonation Both airbreathing and rocket modeof operation of this valved design was discussed Lynch et al70

presented a computational study of an axisymmetric PDE with astraight inlet very similar to that of Eidelman and Grossmann8 andEidelman et al69 Various inlet lengths from 1 to 12 cm and ightMach numbers of 08 and 20 were studied The results were similarto those reported earlier except that some additional details of thedetonation wave front and the air induction system were resolved inthese studies In a related work Lynch and Edelman71 showed thatreductions in scavenging time (and hence overall cycle time) canbe achieved by suitably shaping the inlet

Bussing et al72 compared open-end initiation and closed-end ini-tiation in a simple tube PDE geometry Equivalent thrust productionand fuel ef ciency were observed for the two types of initiation

KAILASANATH 1705

Although a higher head-end pressure was attained during open-endinitiation the pressure continuously dropped for that case unlike theclosed-end initiation where a constant pressure region occurred fora while before reduction of the pressure to the ambient value Morerecently an experimental study has con rmed the differences in theobserved pressure pro les and it has also been concluded that themaximum level of impulse attained is independent of the locationof the detonation initiator73

In the past few years there has been a substantial growth in theappearance of papers related to the PDE and complete sessionshave been devoted to the topic at annual AIAAASMESAEASEEjoint propulsion conferences Because of space considerations onlyselected papers from these years are brie y discussed here Sterlinget al74 conducted one-dimensional numerical investigations of aself-aspiratingPDE operating for multiple cyclesA key observationfrom their studies was that only a portion of the detonation tubecan be lled with fresh charge under self-aspirating operation Inspite of this limitation they reported a speci c impulse of 5151 sfor a hydrogenndashair system However they concluded that the idealperformance of such an engine is ldquonear those of other hydrogen-fueledair breathing enginesrdquo74

Several basic shock tube experiments related to the developmentof a hydrogen-fueled PDE were described by Hinkey et al75 Theseinclude the measurement of detonation wave velocities and de a-gration to detonation transition (DDT) lengths for a range of equiv-alence ratios As expected the DDT lengths were found to be toolarge for practical applications even in hydrogenndashoxygen mixturesTherefore traditional techniques such as the use of a Schelkin spiral(see Ref 76) were tried to reduce the transition length A factor of2ndash4 reduction was observed over the range of equivalence ratios in-vestigated This work75 highlights the dif culties involved in obtain-ing a fully developed detonation in a short distance (as assumed inthe conceptual studies and numerical simulations discussed earlier)and questions if the detonations reported in previous experimentalinvestigations161766 were really fully transitioned from de agra-tions The speci c impulses measured in their experiments (about240 s for hydrogenndashoxygen and 1200 s for hydrogenndashair) were alsostated to be in agreement with their analysis

An interesting PDE concept presented by Bussing77 was that of arotary-valved multiple-pulsed detonation engine Here several det-onation chambers are coupled to an air inlet and fuel source usinga rotary valve as suggested in one of the early papers of Nichollset al16 The rotary valves allow the lling of some detonation cham-bers while others are detonating or exhausting Other papers includepresentation of an overall PDE performance model78 and the exper-imental characterization of the detonation properties of some fuelsfor use in PDEs79

The importance of adequately mixing the fuel and oxidizer washighlighted by the experimental investigations of Stanley et al80

who obtained very low sub-CndashJ velocities when injecting the fueland oxidizer at different times and not taking extra effort to en-sure that they were well mixed The use of turbulence producingdevices appeared to improve signi cantly the mixing and the at-tainment of higher velocities but also resulted in signi cant thrustlosses This study also showed that raising the initial pressure wasbene cial The computational studies of Eidelman et al81 showedthat the PDE engine could operate even for a range of transitionaldetonation regimes that produced nonplanar or not fully developeddetonations Aarnio et al82 discussed two failure modes DDT tran-sition failure and premature ignition In both cases the system con-tinued to operate though there would be loss of thrust In Ref 82detailed discussions of the pressure and thrust histories during a5-Hz 20-cycle operation are also provided (Fig 17) The cycle tocycle variations of peak pressure have been attributed to the lowsampling frequency used82 Both a direct measurement as well asthe integration of the pressure histories were used to estimate the av-erage Isp from hydrogenndashair systems to be between 1116 and 1333 sThe effect of the predetonator volume on the reduction of speci cimpulse was also shown to be a critical issue The effects of inletsand nozzles were not considered in this study

A conceptual design of a multitube PDE with a single air inletduct was discussed by Pegg et al83 Time-dependent computational

Fig 17 Pressure history at one location from a PDE operating at 5 Hzfor 20 cycles (from Ref 82)

uid dynamics analysis indicated that the inlet isolaterdiffusercon-cept would work and did not allow enough time for the formationof destabilizing hammer shocks Performance analysis includingcomponent ef ciencies reported in the paper indicate that opera-tional frequencies of the order of 75ndash100 Hz are required for theMach 12ndash3 ights considered Experimental data indicating mul-ticycle operation at 100 Hz with a rich ethylenendashoxygen mixturewere presented by Sterling et al84 DDT enhancement devices andpredetonators were used in the experiments In spite of this det-onation waves did not occur all of the time and when detonationwaves failed the tube became very hot and testing was suspendedTherefore the maximum testing time was reported to be 05 s Inspite of this limitation the two studies8384 taken together high-light the possibility of PDEs burning hydrocarbon fuels for realisticmissions

Five different nozzle shapes and their effect on performance werestudied computationally by Cambier and Tegner85 Their results in-dicate that the presence of a nozzle can affect the performance ofthe PDE by increasing the thrust delivery during the ignition phaseThe bell-shaped nozzles appeared to give higher performance thanshapes with positive curvature Their results also showed that noz-zles also affected the ow dynamics and hence the timing of thevarious phases of the engine cycle In an experimental study of aconical exhaust nozzle the authors of Ref 86 noted improved per-formance (based on higher velocities which were still sub-CndashJ) butdid not see any effect on the blowdown process More recently87

the effectof various nozzle shapes including converging divergingand straight have been reexamined computationally The converg-ing sections of nozzles introduced shock wave re ections whereasdiverging sections generated a negative thrust for a portion of thecycle due to overexpansion In spite of these limitations the overallconclusion of these studies was that ldquonozzles can drastically in-crease ef ciency of the PDEsrdquo87 There is also some experimentalevidence suggesting that propulsive performances can be increasedby the addition of a nozzle73 However factors such as the effect ofthe nozzles on detonation transmission and the detailed dynamicsof the ow have not yet been elucidated This appears to be an areathat needs further investigation

Other recent efforts have focused on demonstrating the opera-tion of single- and multitube combustors coupled to an inlet using arotary-valvemechanism88 The valve servesto both meter the air owinto the combustor and to isolate the inlet from the high pressuresproduced during the detonation cycle Utilizing multiple combus-tors that ll and detonate out of phase allows the continual use of thein owing air Firing rates of up to 12 Hz per combustor were demon-strated for a hydrogen-fueled system Several other issues related tothe practical development of PDEs have been discussed by Bussinget al89 According to them three of the more important engineer-ing issues are the development of a valve-injection subsystem anignition system and a low-volume predetonator

The use of PDEs for rockets has been revived recently90 Therocket mode of operation is very similar to the description providedearlier of the airbreathing engine with ignition at the closed endexcept that the oxidizer also needs to be injected into the system pe-riodically An advantage of PDEs for rockets would be their higherpower density thus enabling the development of more compact

1706 KAILASANATH

rockets Reference 90 shows the pressure traces from a pulsed det-onation rocket engine operating at 145 Hz on a hydrogenndashoxygenmixture

More recently a two-tube rotary-valved PDE with ight-sizecomponents was operated at 40 Hz per combustor for 30 s withan ethylenendashair mixture91 In addition throttling of the device al-lowed operation at three thrust levels during a 10-s demonstrationexperiment Contrary to some earlier estimates the thermal loadswere signi cant and it was not possible to operate for more than3 s without water cooling It is not clear if this was because ofperiodic detonation failures and consequent de agrative modes ofcombustion as reported in some earlier studies

The work reported in the open literature and discussed earlier hasfocused on gaseous fuels However for volume-limited propulsionapplications PDEs operating on liquid fuels need to be demon-strated In a recent study Brophy et al92 reported on experimentsusing JP-10oxygen and JP-10air aerosols The fuelndashoxygen mix-ture was successfully detonated to obtain an engine operating at5 Hz but the tests with the fuelndashair mixture were not successful

Computational studies continue to make progress in the study ofthe PDE8793 iexcl 97 The effect of nozzle shapes and incomplete transi-tion to CndashJ detonations have been studied by Eidelman and Yang87

They nd that the cycle ef ciency will be virtually the same whethera CndashJ detonation wave is initiated instantaneously or if the CndashJ valueis attainedonly at the end of the detonation chamber Primarilybasedon this study87 they come to an interesting conclusion that ldquoinitia-tion energies that are required for PDE operations can be small andcomparable to energies required for initiation of (other) combustionprocessesrdquo However this is yet to be substantiated experimentallySekar et al93 reexamine open- and closed-end initiation and thebasic self-aspirating con guration of Eidelman8 using more recentcomputational tools Injection and mixing issues that have gener-ally been ignored in previous numerical studies are also beginningto be addressed94 Both front-wall and lateral-wall injections werestudied and the time required for fuel injection was estimated Itis suggested that it may be possible to ll only a portion of thechamber during high-frequency operations Recent computationalstudies also suggest that preconditioning the fuelndashair mixture us-ing a shock wave can signi cantly reduce the DDT distance95 Theperformance estimates mentioned throughout the discussion of thePDE show signi cant variation A systematic review of various per-formance estimates and possible reasons for the observed variationhave been presented9697 These numerical studies suggest that thetime history of the back pressure is a crucial parameter affecting theestimated performance

Other Detonation EnginesThere have also been several other interesting conceptual studies

on the applications of detonations to engines Cambier et al98 haveproposed a pulsed-detonation wave augmentation device as part ofa hybrid engine for a single-stage-to-orbit airbreathing hypersonicvehicle Here the PDE is used for both thrust generation as well asfor mixingcombustion augmentation Two speci c con gurationswere investigated numerically

The concept of a detonation internal combustion engine was pro-posed by Loth and Loth99 and Loth100 Here a conventional pistoncylinder con guration is modi ed to incorporate a separate deto-nation chamber which is isolated by a valve from the compressionchamber during fuel injection and discharges tangentially into a vor-texchamber formedby the piston and cylinder at top deadcenterTherapid detonative combustion is followed by rapid dilution throughmixing with air in the vortex chamber to reduce the formation ofNOx and unburned hydrocarbons and to achieve an overall leanmixture The vortex chamber was also designed to store a portion ofthe detonation waversquos kinetic energy Overall thermal ef cienciessomewhat better than constant volume combustion were postulatedunder ideal conditions

ConclusionsDetonation waves have been explored extensively for propulsion

applications because of their inherent theoretical advantage overde agrative combustion However practical developments to date

have been of idealized systems or laboratory-scale devices or bothThe basic ideas behind most of the current systems being investi-gated have been known for a while but still many of the details needto be understood Advances in computations and experimental di-agnostics appear to be poised to make practical propulsion devicesbased on detonation waves a reality

Two major systems currently under development are the ram ac-celerator and the PDE In both cases the theoretical and computa-tional studies have been far more encouraging than the experimentsHowever in both cases actual demonstrations of devices workingon gaseous fuels have taken place In the case of the ram accelera-tor fuelndashoxygenndashdiluent mixtures have been used and the primaryproblem appears to be heat transfer and related material failuresThe use of gaseous fuels is not a major issue here but new materialsfor the projectiles need to be explored

In the case of the PDE most of the successful operations havebeen with fuelndashoxygen mixtures There still has not been a directcomparison between detailed experimental data and related theoret-ical and computational results Furthermore most of the PDE andrelated work to date have focused on gaseous mixtures PDEs op-erating on multiphase mixtures need to be emphasized because itwould be more practical to use liquid fuels for most of the propul-sion applicationsproposed for these systemsThere are several othergeneral issues that need to be resolved Clearly the evacuation ofexhaust gases is important both from the point of view of controllingthe cycle time and for preventing autoignition and mixing with freshgases The heat transfer and noise from PDEs have also not beenreported in any detail Other issues include inletndashcombustor andcombustorndashnozzle interactions and scalability of the entire device

AcknowledgmentsThis work has been sponsored by the Of ce of Naval Research

through the Mechanics and Energy Conversion Division and theUS Naval Research Laboratory

References1Fickett W and Davis W C Detonation Univ of California Press

Berkeley CA 19792Hoffmann N ldquoReaction Propulsion by Intermittent Detonative Com-

bustionrdquo German Ministry of Supply AI152365 Volkenrode Translation1940 (cited in Ref 16)

3Roy M ldquoPropulsion par Statoreacteur a Detonationrdquo Comptes-Rendusde lrsquoAcademie des Sciences Vol 222 1946 pp 31 32

4Dabora E K and Broda J C ldquoStanding Normal Detonations andOblique Detonations for Propulsionrdquo AIAA Paper 93-2325 July 1993

5Dabora E K ldquoStatus of Gaseous Detonation Waves and Their Rolein Propulsionrdquo Fall Technical Meeting of the Eastern States Section ofthe Combustion Institute Combustion Inst Pittsburgh PA 1994 pp 11ndash

186Rubins P M and Bauer R C ldquoReview of Shock-Induced Supersonic

Combustion Research and Hypersonic Applicationsrdquo Journal of Propulsionand Power Vol 10 No 5 1994 pp 593ndash601

7Eidelman S Grossmann W and Lottati I ldquoReview of PropulsionApplications and Numerical Simulations of the Pulsed Detonation EngineConceptrdquo Journal of Propulsion and Power Vol 7 No 6 1991 pp 857ndash

865 also AIAA Paper 89-2446 July 19898Eidelman S and Grossmann W ldquoPulsed Detonation Engine Experi-

mental and Theoretical Reviewrdquo AIAA Paper 92-3168 July 19929Reingold L ldquoRecherches sur les combustions permanentes apportees

aux foyers a circulation interne supersoniquerdquo ONERA TN2 Dept of En-ergy and Propulsion Study 728-E Paris March 1950

10Bitondo D and Bollay W ldquoPreliminary Performance Analysis ofthe Pulse-Detonation-Jet Engine Systemrdquo Aerophysics Development CorpRept ADC-102-1 April 1952

11Gross R A ldquoExploratory Study of Combustion in Supersonic FlowrdquoUS Air Force Of ce of Scienti c Research TN 59-587 Washington DCJune 1959

12Gross R A and Chinitz W ldquoA Study of Supersonic CombustionrdquoJournal of Aerospace Science Vol 27 No 7 1960 pp 517ndash525

13Nicholls J A Dabora E K and Gealer R L ldquoStudies in Con-nection with Stabilized Gaseous Detonation Wavesrdquo Seventh Symposium(International) on Combustion Combustion Inst Pittsburgh PA 1959 pp766ndash772

14Nicholls J A and Dabora E K ldquoRecent Results on Standing Detona-tion Wavesrdquo Eighth Symposium (International) on Combustion CombustionInst Pittsburgh PA 1962 pp 644ndash655

KAILASANATH 1707

15Dunlap R Brehm R L and Nicholls J A ldquoA Preliminary Studyof the Application of Steady-State Detonative Combustion to a ReactionEnginerdquo Jet Propulsion Vol 28 No 7 1958 pp 451ndash456

16Nicholls J A Wilkinson H R and Morrison R B ldquoIntermittentDetonation as a Thrust-Producing Mechanismrdquo Jet Propulsion Vol 27No 5 1957 pp 534ndash541

17Krzycki L J ldquoPerformance Characteristics of an Intermittent Detona-tion Devicerdquo US Naval Ordnance Test Station US Naval Weapons Rept7655 China Lake CA 1962

18Nicholls J A Cullen R E and Ragland K W ldquoFeasibility Studiesof a Rotating Detonation Wave Rocket Motorrdquo Journal of Spacecraft andRockets Vol 3 No 6 1966 pp 893ndash898

19Adamson T C and Olsson G R ldquoPerformance Analysis of a RotatingDetonation Wave Rocket Enginerdquo Astronautica Acta Vol 13 No 4 1967pp 405ndash415

20Shen P I and Adamson T C ldquoTheoretical Analysis of a RotatingTwo-Phase Detonation in Liquid Rocket Motorsrdquo Astronautica Acta Vol17 No 4 1972 pp 715ndash728

21Clayton R M and Rogero R S ldquoExperimental Measurements on aRotating Detonation-Like Wave Observed During Liquid Rocket ResonantCombustionrdquo Jet Propulsion Lab TR 32-788 California Inst of Technol-ogy Pasadena CA Aug 1965

22Back L H ldquoApplication of Blast Wave Theory to Explosive Propul-sionrdquo Acta Astronautica Vol 2 1975 pp 391ndash407

23Varsi G Back L H and Kim K ldquoBlast Wave in a Nozzle for Propul-sion Applicationsrdquo Acta Astronautica Vol 3 1976 pp 141ndash156

24Kim K Varsi G and Back LH ldquoBlast Wave Analysis for DetonationPropulsionrdquo AIAA Journal Vol 10 No 10 1977 pp 1500ndash1502

25Back L H Dowler W L and Varsi G ldquoDetonation PropulsionExperiments and Theoryrdquo AIAA Journal Vol 21 No 10 1983 pp 1418ndash

142726Kantrowitz A ldquoPropulsion to Orbit by Ground Based Lasersrdquo Astro-

nautics and Aeronautics Vol 10 No 5 1972 p 7427Hyde R A ldquoOne-D Modelling of a Two-Pulse LSD Thrusterrdquo Pro-

ceedings of the 1986 SDIODARPA Workshop on Laser Propulsion editedby J T Kare Lawrence Livermore National Labs 1987 pp 79ndash88

28Emery M H Kailasanath K Oran E S and Gardner J H ldquoCompu-tational Studies of Laser Supported Detonationsrdquo Proceedings of the 1987SDIO Workshop on Laser Propulsion edited by J T Kare Lawrence Liv-ermore National Labs 1990 pp 43ndash55

29Emery M H Kailasanath K and Oran E S ldquoNumerical Simu-lations of Laser Supported Detonationsrdquo 1988 SDIO Workshop on LaserPropulsion Lawrence Livermore National Labs April 1988

30Carrier G F Fendell F McGregor D Cook S and Vazirani MldquoLaser-Initiated Conical Detonation Wave for Supersonic CombustionrdquoAIAA Paper 91-0578 Jan 1991

31Fendell F E Mitchell J McGregor R Magiawala K and Shef eldM ldquoLaser-Initiated Conical Detonation Wave for Supersonic CombustionIIrdquo AIAA Paper 92-0088 Jan 1992

32Carrier G F Fendell F E and Chou M S ldquoLaser-Initiated ConicalDetonation Wave for Supersonic Combustion IIIrdquo AIAA Paper 92-3247July 1992

33Menees G P Adelman H G Cambier J-L and Bowles J V ldquoWaveCombustors for Trans-Atmospheric Vehiclesrdquo Journal of Propulsion andPower Vol 8 No 3 1992 pp 709ndash713

34Cambier J L Adelman H and Menees G ldquoNumerical Simulationsof an Oblique Detonation Wave Enginerdquo AIAA Paper 88-0063 1988

35Atamanchuk T and Sislian J ldquoOn- and Off-design Performance Anal-ysis of Hypersonic Detonation Wave Ramjetsrdquo AIAA Paper 90-2473 1990

36Kuznetsov M M Neyland V J and Sayapin G N ldquoEf ciency In-vestigation of Scramjet with Detonation and Shockless Supersonic Combus-tionrdquo Utchenye Zapisky TsAGI Vol 23 No 2 1992 pp 30ndash37

37Bruckner A P and Hertzberg A ldquoRam Accelerator Direct LaunchSystem for Space Cargordquo International Astronautical Federation Paper 87-211 1987

38Hertzberg A Bruckner A P and Bogdanoff D W ldquoRam Accel-erator A New Chemical Method for Accelerating Projectiles to UltrahighVelocitiesrdquo AIAA Journal Vol 26 No 2 1988 pp 195ndash203

39Kull A Burnham E Knowlen C Bruckner A P and HertzbergA ldquoExperimental Studies of the Superdetonative Ram Accelerator ModesrdquoAIAA Paper 89-2632 1989

40Patz G Seiler F Smeets G and Srulijes J ldquoStatus of ISLrsquos RA-MAC30 with Fin Guided Projectiles Accelerated in a Smooth Borerdquo Pro-ceedings of the Second International Workshop on Ram Accelerators Univof Washington Seattle WA 1995

41Seiler F Patz G Smeets G and Srulijes J ldquoGasdynamic Limitsof Ignition and Combustion of a Gas Mixture in ISLrsquos RAMAC30 ScramAcceleratorrdquo Proceedings of the Second International Workshop on RamAccelerators Univ of Washington Seattle WA 1995

42Seiler F Patz G Smeets G and Srulijes J ldquoIn uence of ProjectileMaterial and Gas Composition on Superdetonative Combustion in ISLrsquosRAMAC30rdquo AIAA Paper 98-3445 July 1998

43Brackett D C and Bogdanoff D W ldquoComputational Investigation ofOblique Detonation Ramjet-in-Tube Conceptsrdquo Journal of Propulsion andPower Vol 5 No 3 1989 pp 276ndash281

44Yungster S Eberhardt S and Bruckner A P ldquoNumerical Simula-tion of Shock-Induced Combustion Generated by High-Speed Projectiles inDetonable Gas Mixturesrdquo AIAA Paper 89-0673 1989

45Rom J and Kivity Y ldquoAccelerating Projectiles Up to 12 kms Utiliz-ing the Continuous Detonation Propulsion Methodrdquo AIAA Paper 88-29691988

46Humphrey J W ldquoParametric Study of an ODW Scramaccelerator forHypersonic Test Facilitiesrdquo AIAA Paper 90-2470 1990

47Yungster S ldquoNavierndashStokes Simulation of the Supersonic CombustionFlow eld in a Ram Acceleratorrdquo AIAA Paper 91-1916 1991

48Li C Kailasanath K Oran E S Boris J P Paper and LandsbergA M ldquoNumerical Simulations of Transient Flows in Ram AcceleratorsrdquoAIAA Paper 93-1916 1993

49Li C Kailasanath K Oran E S Landsberg A M and Boris J PldquoDynamics of Oblique Detonations in Ram Acceleratorsrdquo Shock Waves Vol5 June 1995 pp 97ndash101

50Yungster S and Radhakrishnan K ldquoComputational Study of FlowEstablishment in a Ram Acceleratorrdquo AIAA Paper 98-2489 July 1995

51Choi J Y Jeung I-S and Yoon Y ldquoNumerical Study of ScramAccelerator Starting Characteristicsrdquo AIAA Journal Vol 36 No 6 1998pp 1029ndash1038

52Bezgin L Ganzhelo A Gouskov O and Kopchenov V ldquoSome Nu-merical Investigation Results of Shock-Induced Combustionrdquo Proceedingsof the AIAA 8th International Space Planes and Hypersonic Systems andTechnologies Conference AIAA Reston VA 1998 pp 72ndash82

53Pratt D W Humphrey J W and Glenn D E ldquoMorphology of Stand-ing Oblique Detonation Wavesrdquo Journal of Propulsion and Power Vol 7No 5 1991

54Fujiwara T Matsuo A and Nomoto H ldquoTwo-Dimensional Detona-tion Supported by a Blunt Body or a Wedgerdquo AIAA Paper 88-0089 Jan1988

55Fan B C Sichel M and Kauffman C W ldquoAnalysis of ObliqueShock-Detonation Wave Interactions in the Supersonic Flow of a Com-bustible Mediumrdquo AIAA Paper 88-0441 Jan 1988

56Dabora E K Desbordes D Guerraud C and Wagner H GldquoOblique Detonations at Hypersonic Velocitiesrdquo Dynamics of Detonationsand Explosions Detonations Vol 133 Progress in Astronautics and Aero-nautics AIAA Washington DC 1990 pp 187ndash201

57Li C Kailasanath K and Oran E S ldquoStability of Oblique Detona-tions in Ram Acceleratorsrdquo AIAA Paper 92-0089 Jan 1992

58Li C Kailasanath K and Oran E S ldquoStructure of Reaction WavesBehind Oblique Shocksrdquo Dynamic Aspects of Detonation Vol 153 Progressin Astronautics and Aeronautics AIAA Washington DC 1993 pp 231ndash

24059Li C Kailasanath K and Oran E S ldquoDetonation Structures Behind

Oblique Shocksrdquo Physics of Fluids Vol 6 April 1994 pp 1600ndash161160Da Silva L F and Deshaies B ldquoNumerical Study of Some Ignition

Regimes of Combustible Supersonic Flows over a Wedgerdquo Proceedings ofSecond International Workshop on Ram Accelerators Paper 15 Univ ofWashington Seattle WA 1995

61Thaker A A and Chelliah H K ldquoNumerical Prediction of ObliqueDetonation Wave Structures Using Detailed and Reduced Reaction Mecha-nismsrdquo Combustion Theory and Modelling Vol 1 No 4 1997 pp 347ndash376

62Desbordes D Guerraud C and Hamada L ldquoSupersonic H2-AirCombustions Behind Oblique Shock Wavesrdquo 14th International Colloquiumon Dynamics of Explosions and Reactive Systems Paper D28 Univ ofCoimbra Portugal Aug 1993

63Li C Kailasanath K and Oran E S ldquoDetonation Structures Gener-ated by Multiple Shocks on Ram-Accelerator Projectilesrdquo Combustion andFlame Vol 108 1997 pp 173ndash186

64Viguier C and Desbordes D ldquoConditions of Onset and Stabilizationof Oblique Detonation Wavesrdquo Advances in Experimentation and Compu-tation of Detonations edited by G D Roy S H Frolov K Kailasanathand N Smirnov ENAS Moscow 1998 pp 40 41

65Bussing Tand Pappas G ldquoIntroduction toPulseDetonation EnginesrdquoAIAA Paper 94-0263 1994

66Helman D Shreeve R P and Eidelman S ldquoDetonation Pulse En-ginerdquo AIAA Paper 86-1683 June 1986

67Cambier J L and Adelman H G ldquoPreliminary Numerical Simu-lations of a Pulsed Detonation Wave Enginerdquo AIAA Paper 88-2960 July1988

68Eidelman S Grossmann W and Lottati I ldquoAir-Breathing PulsedDetonationEngine Concept A Numerical Studyrdquo AIAA Paper 90-2420 July1990

69Eidelman S Grossmann W and Lottati I ldquoComputational Analysisof Pulsed Detonation Engines and Applicationsrdquo AIAA Paper 90-0460 Jan1990

1704 KAILASANATH

PDEsThe other major current propulsion application involving detona-

tion waves is the PDE The basic concept behind a PDE is simpleDetonation is initiated repeatedly at either the closed or the openend of a detonation chamber that is lled with a premixed fuelndashairmixture Figure 15 (based on Ref 65) illustrates the concept forthe case when the detonation wave is initiated at the closed endAs shown in this schematic a one-dimensional planar detonation isinitiated near the closed end of the detonation chamber and travelsat the CndashJ velocity toward the open end A set of rarefaction wavestrail the detonation wave to reduce the velocity to zero at the closedend of the tube When the planar detonation wave exits the chamberanother set of expansion waves is generated that travels toward theclosed end These waves evacuate the burned detonation productsresulting in a cool empty chamber that is ready to be lled with afresh fuelndashair mixture The entire cycle repeats resulting in a pe-riodic high-pressure zone near the closed end of the chamber Theintegrated effectof this high pressure over the closed end (represent-ing a thrust wall) produces thrust This is of course a very idealizedpicture of the system Before discussing progress made in the re-cent years in developing a PDE a brief review of the background isprovided

The early work on the use of intermittent detonations has alreadybeen discussed In the late 1980s this concept of using intermittentor pulsed detonations was reexamined experimentally at the NavalPostgraduate School66 An ethylenendashoxygen detonation wave in asmall-diameter tube was used as a predetonator to initiate detona-tions in a larger tube containing an ethylenendashair mixture Periodicfuel injection within the naturally aspirated tube resulted in an in-termittent frequency of 25 Hz Speci c impulse estimated usingthe pressure time history and the amount of fuel consumed rangedfrom 1000 to 1400 s The velocities of the observed waves (lessthan 1 kms) are signi cantly below the CndashJ detonation wave ve-locities for the reported mixtures indicating that a fully developeddetonation wave was not formed

The rst purely computational study of the PDE reported in theopen literature appears to be the work of Cambier and Adelman67

The system simulated consisted of a 50-cm-long main tube attachedto a 43-cm-long diverging nozzle Quasi-one-dimensional simula-tions using the Euler equations for unsteady ow were carried outwith a total variation diminishing scheme and multistep nite ratekinetics A detonation was initiated at the closed end of a tube lledwith a premixed stoichiometric hydrogenndashair mixture at 3 atmOverall performance calculated by integrating the instantaneousthrust and the fuel ow rates gave a speci c impulse Isp of about6500 s and a range of operating frequencies up to 667 Hz Issues ofignition energies and transition to detonation were not consideredand the dif culties with attaining rapid mixing of fuel and air wasavoided by using a premixed fuelndashair mixture These assumptionsare still typical of those used in more recent simulations

The rst look at the interactions between the ow inside the cham-ber and that outside appears to be the work of Eidelman et al7

who simulated a cylindrical detonation chamber (about 15-cm diam

Fig 15 Schematic of an idealized PDE showing various stages duringone cycle of operation (adapted from Ref 65)

Fig 16 Generic PDE device used for somecomputational studieswheredetonation is initiated near the open end (from Ref 69)

and 15 cm long) with a small converging nozzle In their two-dimensional simulations a self-similarsolution for a planar detona-tion was imposed near the open end and made to travel toward theclosed end The re ection of the wave from the closed end createda higher pressure at the wall than during initiation near the closedend Hence this mode of initiation was considered to be better Af-ter the detonation left the chamber the pressure in the chamber wasobserved to fall below the ambient resulting in ingestion of outsideair In subsequent works86869 the same numerical techniques andmodels were used to investigate a slightly different con gurationwhere the nozzle at the aft end was removed and an inlet was in-troduced near the closed end This con guration has become oneof the popular systems for two-dimensional (and axisymmetric) nu-merical simulations Again detonation was assumed to be initiatedat the open end and traveled toward the closed end In this con gu-ration when the pressure falls below atmospheric in the detonationchamber air is ingested through the inlets in the front even duringstatic operations Thus a valveless self-aspiratingoperation is pos-sible These simulations also introduced the discussion of open-endinitiation vs closed-end initiation and their relative advantages Thecurrent status of this issue will be discussed later

Many of the early efforts on intermittent or PDEs have been re-viewed in Ref 7 and 8 where a link is also made between thesedetonation engines and pulse jet engines (which do not use detona-tive combustion) In addition Ref 8 provides a detailed summaryof the numerical investigations conducted by the authorsrsquo group ona generic PDE device shown in Fig 16 A characteristic feature ofthis device is that the detonation wave is initiated at the open endand travels toward a thrust wall at the closed end An advantage ofthis approach is that it is valveless and self-aspirating Air is en-trained into the chamber through inlets near the closed end whenthe pressure falls below atmospheric pressure during part of the cy-cle Unsteady Euler simulations were used to study the operationof the device from static (M = 0) to supersonic (M = 2) conditionsAs stated in Ref 8 a proof of principle experimental demonstra-tion of the PDE mode of operation where detonation is initiated atthe open end still needed to be done Furthermore issues such asdetonation initiation and structure and fuelndashair injection and mixingalso needed to be studied

Bussing and Pappas65 provided a detailed description of thebasic operation of an ideal zed PDE They also reported someone-dimensional studies of PDEs burning hydrogenndashoxygen andhydrogenndashair mixtures The detonation wave was initiated at theclosedend in this engine using a high-temperature and high-pressureregion An advantage of this design is the con nement provided atthe closed end which should enable a more rapid initiation andestablishment of a detonation Both airbreathing and rocket modeof operation of this valved design was discussed Lynch et al70

presented a computational study of an axisymmetric PDE with astraight inlet very similar to that of Eidelman and Grossmann8 andEidelman et al69 Various inlet lengths from 1 to 12 cm and ightMach numbers of 08 and 20 were studied The results were similarto those reported earlier except that some additional details of thedetonation wave front and the air induction system were resolved inthese studies In a related work Lynch and Edelman71 showed thatreductions in scavenging time (and hence overall cycle time) canbe achieved by suitably shaping the inlet

Bussing et al72 compared open-end initiation and closed-end ini-tiation in a simple tube PDE geometry Equivalent thrust productionand fuel ef ciency were observed for the two types of initiation

KAILASANATH 1705

Although a higher head-end pressure was attained during open-endinitiation the pressure continuously dropped for that case unlike theclosed-end initiation where a constant pressure region occurred fora while before reduction of the pressure to the ambient value Morerecently an experimental study has con rmed the differences in theobserved pressure pro les and it has also been concluded that themaximum level of impulse attained is independent of the locationof the detonation initiator73

In the past few years there has been a substantial growth in theappearance of papers related to the PDE and complete sessionshave been devoted to the topic at annual AIAAASMESAEASEEjoint propulsion conferences Because of space considerations onlyselected papers from these years are brie y discussed here Sterlinget al74 conducted one-dimensional numerical investigations of aself-aspiratingPDE operating for multiple cyclesA key observationfrom their studies was that only a portion of the detonation tubecan be lled with fresh charge under self-aspirating operation Inspite of this limitation they reported a speci c impulse of 5151 sfor a hydrogenndashair system However they concluded that the idealperformance of such an engine is ldquonear those of other hydrogen-fueledair breathing enginesrdquo74

Several basic shock tube experiments related to the developmentof a hydrogen-fueled PDE were described by Hinkey et al75 Theseinclude the measurement of detonation wave velocities and de a-gration to detonation transition (DDT) lengths for a range of equiv-alence ratios As expected the DDT lengths were found to be toolarge for practical applications even in hydrogenndashoxygen mixturesTherefore traditional techniques such as the use of a Schelkin spiral(see Ref 76) were tried to reduce the transition length A factor of2ndash4 reduction was observed over the range of equivalence ratios in-vestigated This work75 highlights the dif culties involved in obtain-ing a fully developed detonation in a short distance (as assumed inthe conceptual studies and numerical simulations discussed earlier)and questions if the detonations reported in previous experimentalinvestigations161766 were really fully transitioned from de agra-tions The speci c impulses measured in their experiments (about240 s for hydrogenndashoxygen and 1200 s for hydrogenndashair) were alsostated to be in agreement with their analysis

An interesting PDE concept presented by Bussing77 was that of arotary-valved multiple-pulsed detonation engine Here several det-onation chambers are coupled to an air inlet and fuel source usinga rotary valve as suggested in one of the early papers of Nichollset al16 The rotary valves allow the lling of some detonation cham-bers while others are detonating or exhausting Other papers includepresentation of an overall PDE performance model78 and the exper-imental characterization of the detonation properties of some fuelsfor use in PDEs79

The importance of adequately mixing the fuel and oxidizer washighlighted by the experimental investigations of Stanley et al80

who obtained very low sub-CndashJ velocities when injecting the fueland oxidizer at different times and not taking extra effort to en-sure that they were well mixed The use of turbulence producingdevices appeared to improve signi cantly the mixing and the at-tainment of higher velocities but also resulted in signi cant thrustlosses This study also showed that raising the initial pressure wasbene cial The computational studies of Eidelman et al81 showedthat the PDE engine could operate even for a range of transitionaldetonation regimes that produced nonplanar or not fully developeddetonations Aarnio et al82 discussed two failure modes DDT tran-sition failure and premature ignition In both cases the system con-tinued to operate though there would be loss of thrust In Ref 82detailed discussions of the pressure and thrust histories during a5-Hz 20-cycle operation are also provided (Fig 17) The cycle tocycle variations of peak pressure have been attributed to the lowsampling frequency used82 Both a direct measurement as well asthe integration of the pressure histories were used to estimate the av-erage Isp from hydrogenndashair systems to be between 1116 and 1333 sThe effect of the predetonator volume on the reduction of speci cimpulse was also shown to be a critical issue The effects of inletsand nozzles were not considered in this study

A conceptual design of a multitube PDE with a single air inletduct was discussed by Pegg et al83 Time-dependent computational

Fig 17 Pressure history at one location from a PDE operating at 5 Hzfor 20 cycles (from Ref 82)

uid dynamics analysis indicated that the inlet isolaterdiffusercon-cept would work and did not allow enough time for the formationof destabilizing hammer shocks Performance analysis includingcomponent ef ciencies reported in the paper indicate that opera-tional frequencies of the order of 75ndash100 Hz are required for theMach 12ndash3 ights considered Experimental data indicating mul-ticycle operation at 100 Hz with a rich ethylenendashoxygen mixturewere presented by Sterling et al84 DDT enhancement devices andpredetonators were used in the experiments In spite of this det-onation waves did not occur all of the time and when detonationwaves failed the tube became very hot and testing was suspendedTherefore the maximum testing time was reported to be 05 s Inspite of this limitation the two studies8384 taken together high-light the possibility of PDEs burning hydrocarbon fuels for realisticmissions

Five different nozzle shapes and their effect on performance werestudied computationally by Cambier and Tegner85 Their results in-dicate that the presence of a nozzle can affect the performance ofthe PDE by increasing the thrust delivery during the ignition phaseThe bell-shaped nozzles appeared to give higher performance thanshapes with positive curvature Their results also showed that noz-zles also affected the ow dynamics and hence the timing of thevarious phases of the engine cycle In an experimental study of aconical exhaust nozzle the authors of Ref 86 noted improved per-formance (based on higher velocities which were still sub-CndashJ) butdid not see any effect on the blowdown process More recently87

the effectof various nozzle shapes including converging divergingand straight have been reexamined computationally The converg-ing sections of nozzles introduced shock wave re ections whereasdiverging sections generated a negative thrust for a portion of thecycle due to overexpansion In spite of these limitations the overallconclusion of these studies was that ldquonozzles can drastically in-crease ef ciency of the PDEsrdquo87 There is also some experimentalevidence suggesting that propulsive performances can be increasedby the addition of a nozzle73 However factors such as the effect ofthe nozzles on detonation transmission and the detailed dynamicsof the ow have not yet been elucidated This appears to be an areathat needs further investigation

Other recent efforts have focused on demonstrating the opera-tion of single- and multitube combustors coupled to an inlet using arotary-valvemechanism88 The valve servesto both meter the air owinto the combustor and to isolate the inlet from the high pressuresproduced during the detonation cycle Utilizing multiple combus-tors that ll and detonate out of phase allows the continual use of thein owing air Firing rates of up to 12 Hz per combustor were demon-strated for a hydrogen-fueled system Several other issues related tothe practical development of PDEs have been discussed by Bussinget al89 According to them three of the more important engineer-ing issues are the development of a valve-injection subsystem anignition system and a low-volume predetonator

The use of PDEs for rockets has been revived recently90 Therocket mode of operation is very similar to the description providedearlier of the airbreathing engine with ignition at the closed endexcept that the oxidizer also needs to be injected into the system pe-riodically An advantage of PDEs for rockets would be their higherpower density thus enabling the development of more compact

1706 KAILASANATH

rockets Reference 90 shows the pressure traces from a pulsed det-onation rocket engine operating at 145 Hz on a hydrogenndashoxygenmixture

More recently a two-tube rotary-valved PDE with ight-sizecomponents was operated at 40 Hz per combustor for 30 s withan ethylenendashair mixture91 In addition throttling of the device al-lowed operation at three thrust levels during a 10-s demonstrationexperiment Contrary to some earlier estimates the thermal loadswere signi cant and it was not possible to operate for more than3 s without water cooling It is not clear if this was because ofperiodic detonation failures and consequent de agrative modes ofcombustion as reported in some earlier studies

The work reported in the open literature and discussed earlier hasfocused on gaseous fuels However for volume-limited propulsionapplications PDEs operating on liquid fuels need to be demon-strated In a recent study Brophy et al92 reported on experimentsusing JP-10oxygen and JP-10air aerosols The fuelndashoxygen mix-ture was successfully detonated to obtain an engine operating at5 Hz but the tests with the fuelndashair mixture were not successful

Computational studies continue to make progress in the study ofthe PDE8793 iexcl 97 The effect of nozzle shapes and incomplete transi-tion to CndashJ detonations have been studied by Eidelman and Yang87

They nd that the cycle ef ciency will be virtually the same whethera CndashJ detonation wave is initiated instantaneously or if the CndashJ valueis attainedonly at the end of the detonation chamber Primarilybasedon this study87 they come to an interesting conclusion that ldquoinitia-tion energies that are required for PDE operations can be small andcomparable to energies required for initiation of (other) combustionprocessesrdquo However this is yet to be substantiated experimentallySekar et al93 reexamine open- and closed-end initiation and thebasic self-aspirating con guration of Eidelman8 using more recentcomputational tools Injection and mixing issues that have gener-ally been ignored in previous numerical studies are also beginningto be addressed94 Both front-wall and lateral-wall injections werestudied and the time required for fuel injection was estimated Itis suggested that it may be possible to ll only a portion of thechamber during high-frequency operations Recent computationalstudies also suggest that preconditioning the fuelndashair mixture us-ing a shock wave can signi cantly reduce the DDT distance95 Theperformance estimates mentioned throughout the discussion of thePDE show signi cant variation A systematic review of various per-formance estimates and possible reasons for the observed variationhave been presented9697 These numerical studies suggest that thetime history of the back pressure is a crucial parameter affecting theestimated performance

Other Detonation EnginesThere have also been several other interesting conceptual studies

on the applications of detonations to engines Cambier et al98 haveproposed a pulsed-detonation wave augmentation device as part ofa hybrid engine for a single-stage-to-orbit airbreathing hypersonicvehicle Here the PDE is used for both thrust generation as well asfor mixingcombustion augmentation Two speci c con gurationswere investigated numerically

The concept of a detonation internal combustion engine was pro-posed by Loth and Loth99 and Loth100 Here a conventional pistoncylinder con guration is modi ed to incorporate a separate deto-nation chamber which is isolated by a valve from the compressionchamber during fuel injection and discharges tangentially into a vor-texchamber formedby the piston and cylinder at top deadcenterTherapid detonative combustion is followed by rapid dilution throughmixing with air in the vortex chamber to reduce the formation ofNOx and unburned hydrocarbons and to achieve an overall leanmixture The vortex chamber was also designed to store a portion ofthe detonation waversquos kinetic energy Overall thermal ef cienciessomewhat better than constant volume combustion were postulatedunder ideal conditions

ConclusionsDetonation waves have been explored extensively for propulsion

applications because of their inherent theoretical advantage overde agrative combustion However practical developments to date

have been of idealized systems or laboratory-scale devices or bothThe basic ideas behind most of the current systems being investi-gated have been known for a while but still many of the details needto be understood Advances in computations and experimental di-agnostics appear to be poised to make practical propulsion devicesbased on detonation waves a reality

Two major systems currently under development are the ram ac-celerator and the PDE In both cases the theoretical and computa-tional studies have been far more encouraging than the experimentsHowever in both cases actual demonstrations of devices workingon gaseous fuels have taken place In the case of the ram accelera-tor fuelndashoxygenndashdiluent mixtures have been used and the primaryproblem appears to be heat transfer and related material failuresThe use of gaseous fuels is not a major issue here but new materialsfor the projectiles need to be explored

In the case of the PDE most of the successful operations havebeen with fuelndashoxygen mixtures There still has not been a directcomparison between detailed experimental data and related theoret-ical and computational results Furthermore most of the PDE andrelated work to date have focused on gaseous mixtures PDEs op-erating on multiphase mixtures need to be emphasized because itwould be more practical to use liquid fuels for most of the propul-sion applicationsproposed for these systemsThere are several othergeneral issues that need to be resolved Clearly the evacuation ofexhaust gases is important both from the point of view of controllingthe cycle time and for preventing autoignition and mixing with freshgases The heat transfer and noise from PDEs have also not beenreported in any detail Other issues include inletndashcombustor andcombustorndashnozzle interactions and scalability of the entire device

AcknowledgmentsThis work has been sponsored by the Of ce of Naval Research

through the Mechanics and Energy Conversion Division and theUS Naval Research Laboratory

References1Fickett W and Davis W C Detonation Univ of California Press

Berkeley CA 19792Hoffmann N ldquoReaction Propulsion by Intermittent Detonative Com-

bustionrdquo German Ministry of Supply AI152365 Volkenrode Translation1940 (cited in Ref 16)

3Roy M ldquoPropulsion par Statoreacteur a Detonationrdquo Comptes-Rendusde lrsquoAcademie des Sciences Vol 222 1946 pp 31 32

4Dabora E K and Broda J C ldquoStanding Normal Detonations andOblique Detonations for Propulsionrdquo AIAA Paper 93-2325 July 1993

5Dabora E K ldquoStatus of Gaseous Detonation Waves and Their Rolein Propulsionrdquo Fall Technical Meeting of the Eastern States Section ofthe Combustion Institute Combustion Inst Pittsburgh PA 1994 pp 11ndash

186Rubins P M and Bauer R C ldquoReview of Shock-Induced Supersonic

Combustion Research and Hypersonic Applicationsrdquo Journal of Propulsionand Power Vol 10 No 5 1994 pp 593ndash601

7Eidelman S Grossmann W and Lottati I ldquoReview of PropulsionApplications and Numerical Simulations of the Pulsed Detonation EngineConceptrdquo Journal of Propulsion and Power Vol 7 No 6 1991 pp 857ndash

865 also AIAA Paper 89-2446 July 19898Eidelman S and Grossmann W ldquoPulsed Detonation Engine Experi-

mental and Theoretical Reviewrdquo AIAA Paper 92-3168 July 19929Reingold L ldquoRecherches sur les combustions permanentes apportees

aux foyers a circulation interne supersoniquerdquo ONERA TN2 Dept of En-ergy and Propulsion Study 728-E Paris March 1950

10Bitondo D and Bollay W ldquoPreliminary Performance Analysis ofthe Pulse-Detonation-Jet Engine Systemrdquo Aerophysics Development CorpRept ADC-102-1 April 1952

11Gross R A ldquoExploratory Study of Combustion in Supersonic FlowrdquoUS Air Force Of ce of Scienti c Research TN 59-587 Washington DCJune 1959

12Gross R A and Chinitz W ldquoA Study of Supersonic CombustionrdquoJournal of Aerospace Science Vol 27 No 7 1960 pp 517ndash525

13Nicholls J A Dabora E K and Gealer R L ldquoStudies in Con-nection with Stabilized Gaseous Detonation Wavesrdquo Seventh Symposium(International) on Combustion Combustion Inst Pittsburgh PA 1959 pp766ndash772

14Nicholls J A and Dabora E K ldquoRecent Results on Standing Detona-tion Wavesrdquo Eighth Symposium (International) on Combustion CombustionInst Pittsburgh PA 1962 pp 644ndash655

KAILASANATH 1707

15Dunlap R Brehm R L and Nicholls J A ldquoA Preliminary Studyof the Application of Steady-State Detonative Combustion to a ReactionEnginerdquo Jet Propulsion Vol 28 No 7 1958 pp 451ndash456

16Nicholls J A Wilkinson H R and Morrison R B ldquoIntermittentDetonation as a Thrust-Producing Mechanismrdquo Jet Propulsion Vol 27No 5 1957 pp 534ndash541

17Krzycki L J ldquoPerformance Characteristics of an Intermittent Detona-tion Devicerdquo US Naval Ordnance Test Station US Naval Weapons Rept7655 China Lake CA 1962

18Nicholls J A Cullen R E and Ragland K W ldquoFeasibility Studiesof a Rotating Detonation Wave Rocket Motorrdquo Journal of Spacecraft andRockets Vol 3 No 6 1966 pp 893ndash898

19Adamson T C and Olsson G R ldquoPerformance Analysis of a RotatingDetonation Wave Rocket Enginerdquo Astronautica Acta Vol 13 No 4 1967pp 405ndash415

20Shen P I and Adamson T C ldquoTheoretical Analysis of a RotatingTwo-Phase Detonation in Liquid Rocket Motorsrdquo Astronautica Acta Vol17 No 4 1972 pp 715ndash728

21Clayton R M and Rogero R S ldquoExperimental Measurements on aRotating Detonation-Like Wave Observed During Liquid Rocket ResonantCombustionrdquo Jet Propulsion Lab TR 32-788 California Inst of Technol-ogy Pasadena CA Aug 1965

22Back L H ldquoApplication of Blast Wave Theory to Explosive Propul-sionrdquo Acta Astronautica Vol 2 1975 pp 391ndash407

23Varsi G Back L H and Kim K ldquoBlast Wave in a Nozzle for Propul-sion Applicationsrdquo Acta Astronautica Vol 3 1976 pp 141ndash156

24Kim K Varsi G and Back LH ldquoBlast Wave Analysis for DetonationPropulsionrdquo AIAA Journal Vol 10 No 10 1977 pp 1500ndash1502

25Back L H Dowler W L and Varsi G ldquoDetonation PropulsionExperiments and Theoryrdquo AIAA Journal Vol 21 No 10 1983 pp 1418ndash

142726Kantrowitz A ldquoPropulsion to Orbit by Ground Based Lasersrdquo Astro-

nautics and Aeronautics Vol 10 No 5 1972 p 7427Hyde R A ldquoOne-D Modelling of a Two-Pulse LSD Thrusterrdquo Pro-

ceedings of the 1986 SDIODARPA Workshop on Laser Propulsion editedby J T Kare Lawrence Livermore National Labs 1987 pp 79ndash88

28Emery M H Kailasanath K Oran E S and Gardner J H ldquoCompu-tational Studies of Laser Supported Detonationsrdquo Proceedings of the 1987SDIO Workshop on Laser Propulsion edited by J T Kare Lawrence Liv-ermore National Labs 1990 pp 43ndash55

29Emery M H Kailasanath K and Oran E S ldquoNumerical Simu-lations of Laser Supported Detonationsrdquo 1988 SDIO Workshop on LaserPropulsion Lawrence Livermore National Labs April 1988

30Carrier G F Fendell F McGregor D Cook S and Vazirani MldquoLaser-Initiated Conical Detonation Wave for Supersonic CombustionrdquoAIAA Paper 91-0578 Jan 1991

31Fendell F E Mitchell J McGregor R Magiawala K and Shef eldM ldquoLaser-Initiated Conical Detonation Wave for Supersonic CombustionIIrdquo AIAA Paper 92-0088 Jan 1992

32Carrier G F Fendell F E and Chou M S ldquoLaser-Initiated ConicalDetonation Wave for Supersonic Combustion IIIrdquo AIAA Paper 92-3247July 1992

33Menees G P Adelman H G Cambier J-L and Bowles J V ldquoWaveCombustors for Trans-Atmospheric Vehiclesrdquo Journal of Propulsion andPower Vol 8 No 3 1992 pp 709ndash713

34Cambier J L Adelman H and Menees G ldquoNumerical Simulationsof an Oblique Detonation Wave Enginerdquo AIAA Paper 88-0063 1988

35Atamanchuk T and Sislian J ldquoOn- and Off-design Performance Anal-ysis of Hypersonic Detonation Wave Ramjetsrdquo AIAA Paper 90-2473 1990

36Kuznetsov M M Neyland V J and Sayapin G N ldquoEf ciency In-vestigation of Scramjet with Detonation and Shockless Supersonic Combus-tionrdquo Utchenye Zapisky TsAGI Vol 23 No 2 1992 pp 30ndash37

37Bruckner A P and Hertzberg A ldquoRam Accelerator Direct LaunchSystem for Space Cargordquo International Astronautical Federation Paper 87-211 1987

38Hertzberg A Bruckner A P and Bogdanoff D W ldquoRam Accel-erator A New Chemical Method for Accelerating Projectiles to UltrahighVelocitiesrdquo AIAA Journal Vol 26 No 2 1988 pp 195ndash203

39Kull A Burnham E Knowlen C Bruckner A P and HertzbergA ldquoExperimental Studies of the Superdetonative Ram Accelerator ModesrdquoAIAA Paper 89-2632 1989

40Patz G Seiler F Smeets G and Srulijes J ldquoStatus of ISLrsquos RA-MAC30 with Fin Guided Projectiles Accelerated in a Smooth Borerdquo Pro-ceedings of the Second International Workshop on Ram Accelerators Univof Washington Seattle WA 1995

41Seiler F Patz G Smeets G and Srulijes J ldquoGasdynamic Limitsof Ignition and Combustion of a Gas Mixture in ISLrsquos RAMAC30 ScramAcceleratorrdquo Proceedings of the Second International Workshop on RamAccelerators Univ of Washington Seattle WA 1995

42Seiler F Patz G Smeets G and Srulijes J ldquoIn uence of ProjectileMaterial and Gas Composition on Superdetonative Combustion in ISLrsquosRAMAC30rdquo AIAA Paper 98-3445 July 1998

43Brackett D C and Bogdanoff D W ldquoComputational Investigation ofOblique Detonation Ramjet-in-Tube Conceptsrdquo Journal of Propulsion andPower Vol 5 No 3 1989 pp 276ndash281

44Yungster S Eberhardt S and Bruckner A P ldquoNumerical Simula-tion of Shock-Induced Combustion Generated by High-Speed Projectiles inDetonable Gas Mixturesrdquo AIAA Paper 89-0673 1989

45Rom J and Kivity Y ldquoAccelerating Projectiles Up to 12 kms Utiliz-ing the Continuous Detonation Propulsion Methodrdquo AIAA Paper 88-29691988

46Humphrey J W ldquoParametric Study of an ODW Scramaccelerator forHypersonic Test Facilitiesrdquo AIAA Paper 90-2470 1990

47Yungster S ldquoNavierndashStokes Simulation of the Supersonic CombustionFlow eld in a Ram Acceleratorrdquo AIAA Paper 91-1916 1991

48Li C Kailasanath K Oran E S Boris J P Paper and LandsbergA M ldquoNumerical Simulations of Transient Flows in Ram AcceleratorsrdquoAIAA Paper 93-1916 1993

49Li C Kailasanath K Oran E S Landsberg A M and Boris J PldquoDynamics of Oblique Detonations in Ram Acceleratorsrdquo Shock Waves Vol5 June 1995 pp 97ndash101

50Yungster S and Radhakrishnan K ldquoComputational Study of FlowEstablishment in a Ram Acceleratorrdquo AIAA Paper 98-2489 July 1995

51Choi J Y Jeung I-S and Yoon Y ldquoNumerical Study of ScramAccelerator Starting Characteristicsrdquo AIAA Journal Vol 36 No 6 1998pp 1029ndash1038

52Bezgin L Ganzhelo A Gouskov O and Kopchenov V ldquoSome Nu-merical Investigation Results of Shock-Induced Combustionrdquo Proceedingsof the AIAA 8th International Space Planes and Hypersonic Systems andTechnologies Conference AIAA Reston VA 1998 pp 72ndash82

53Pratt D W Humphrey J W and Glenn D E ldquoMorphology of Stand-ing Oblique Detonation Wavesrdquo Journal of Propulsion and Power Vol 7No 5 1991

54Fujiwara T Matsuo A and Nomoto H ldquoTwo-Dimensional Detona-tion Supported by a Blunt Body or a Wedgerdquo AIAA Paper 88-0089 Jan1988

55Fan B C Sichel M and Kauffman C W ldquoAnalysis of ObliqueShock-Detonation Wave Interactions in the Supersonic Flow of a Com-bustible Mediumrdquo AIAA Paper 88-0441 Jan 1988

56Dabora E K Desbordes D Guerraud C and Wagner H GldquoOblique Detonations at Hypersonic Velocitiesrdquo Dynamics of Detonationsand Explosions Detonations Vol 133 Progress in Astronautics and Aero-nautics AIAA Washington DC 1990 pp 187ndash201

57Li C Kailasanath K and Oran E S ldquoStability of Oblique Detona-tions in Ram Acceleratorsrdquo AIAA Paper 92-0089 Jan 1992

58Li C Kailasanath K and Oran E S ldquoStructure of Reaction WavesBehind Oblique Shocksrdquo Dynamic Aspects of Detonation Vol 153 Progressin Astronautics and Aeronautics AIAA Washington DC 1993 pp 231ndash

24059Li C Kailasanath K and Oran E S ldquoDetonation Structures Behind

Oblique Shocksrdquo Physics of Fluids Vol 6 April 1994 pp 1600ndash161160Da Silva L F and Deshaies B ldquoNumerical Study of Some Ignition

Regimes of Combustible Supersonic Flows over a Wedgerdquo Proceedings ofSecond International Workshop on Ram Accelerators Paper 15 Univ ofWashington Seattle WA 1995

61Thaker A A and Chelliah H K ldquoNumerical Prediction of ObliqueDetonation Wave Structures Using Detailed and Reduced Reaction Mecha-nismsrdquo Combustion Theory and Modelling Vol 1 No 4 1997 pp 347ndash376

62Desbordes D Guerraud C and Hamada L ldquoSupersonic H2-AirCombustions Behind Oblique Shock Wavesrdquo 14th International Colloquiumon Dynamics of Explosions and Reactive Systems Paper D28 Univ ofCoimbra Portugal Aug 1993

63Li C Kailasanath K and Oran E S ldquoDetonation Structures Gener-ated by Multiple Shocks on Ram-Accelerator Projectilesrdquo Combustion andFlame Vol 108 1997 pp 173ndash186

64Viguier C and Desbordes D ldquoConditions of Onset and Stabilizationof Oblique Detonation Wavesrdquo Advances in Experimentation and Compu-tation of Detonations edited by G D Roy S H Frolov K Kailasanathand N Smirnov ENAS Moscow 1998 pp 40 41

65Bussing Tand Pappas G ldquoIntroduction toPulseDetonation EnginesrdquoAIAA Paper 94-0263 1994

66Helman D Shreeve R P and Eidelman S ldquoDetonation Pulse En-ginerdquo AIAA Paper 86-1683 June 1986

67Cambier J L and Adelman H G ldquoPreliminary Numerical Simu-lations of a Pulsed Detonation Wave Enginerdquo AIAA Paper 88-2960 July1988

68Eidelman S Grossmann W and Lottati I ldquoAir-Breathing PulsedDetonationEngine Concept A Numerical Studyrdquo AIAA Paper 90-2420 July1990

69Eidelman S Grossmann W and Lottati I ldquoComputational Analysisof Pulsed Detonation Engines and Applicationsrdquo AIAA Paper 90-0460 Jan1990

KAILASANATH 1705

Although a higher head-end pressure was attained during open-endinitiation the pressure continuously dropped for that case unlike theclosed-end initiation where a constant pressure region occurred fora while before reduction of the pressure to the ambient value Morerecently an experimental study has con rmed the differences in theobserved pressure pro les and it has also been concluded that themaximum level of impulse attained is independent of the locationof the detonation initiator73

In the past few years there has been a substantial growth in theappearance of papers related to the PDE and complete sessionshave been devoted to the topic at annual AIAAASMESAEASEEjoint propulsion conferences Because of space considerations onlyselected papers from these years are brie y discussed here Sterlinget al74 conducted one-dimensional numerical investigations of aself-aspiratingPDE operating for multiple cyclesA key observationfrom their studies was that only a portion of the detonation tubecan be lled with fresh charge under self-aspirating operation Inspite of this limitation they reported a speci c impulse of 5151 sfor a hydrogenndashair system However they concluded that the idealperformance of such an engine is ldquonear those of other hydrogen-fueledair breathing enginesrdquo74

Several basic shock tube experiments related to the developmentof a hydrogen-fueled PDE were described by Hinkey et al75 Theseinclude the measurement of detonation wave velocities and de a-gration to detonation transition (DDT) lengths for a range of equiv-alence ratios As expected the DDT lengths were found to be toolarge for practical applications even in hydrogenndashoxygen mixturesTherefore traditional techniques such as the use of a Schelkin spiral(see Ref 76) were tried to reduce the transition length A factor of2ndash4 reduction was observed over the range of equivalence ratios in-vestigated This work75 highlights the dif culties involved in obtain-ing a fully developed detonation in a short distance (as assumed inthe conceptual studies and numerical simulations discussed earlier)and questions if the detonations reported in previous experimentalinvestigations161766 were really fully transitioned from de agra-tions The speci c impulses measured in their experiments (about240 s for hydrogenndashoxygen and 1200 s for hydrogenndashair) were alsostated to be in agreement with their analysis

An interesting PDE concept presented by Bussing77 was that of arotary-valved multiple-pulsed detonation engine Here several det-onation chambers are coupled to an air inlet and fuel source usinga rotary valve as suggested in one of the early papers of Nichollset al16 The rotary valves allow the lling of some detonation cham-bers while others are detonating or exhausting Other papers includepresentation of an overall PDE performance model78 and the exper-imental characterization of the detonation properties of some fuelsfor use in PDEs79

The importance of adequately mixing the fuel and oxidizer washighlighted by the experimental investigations of Stanley et al80

who obtained very low sub-CndashJ velocities when injecting the fueland oxidizer at different times and not taking extra effort to en-sure that they were well mixed The use of turbulence producingdevices appeared to improve signi cantly the mixing and the at-tainment of higher velocities but also resulted in signi cant thrustlosses This study also showed that raising the initial pressure wasbene cial The computational studies of Eidelman et al81 showedthat the PDE engine could operate even for a range of transitionaldetonation regimes that produced nonplanar or not fully developeddetonations Aarnio et al82 discussed two failure modes DDT tran-sition failure and premature ignition In both cases the system con-tinued to operate though there would be loss of thrust In Ref 82detailed discussions of the pressure and thrust histories during a5-Hz 20-cycle operation are also provided (Fig 17) The cycle tocycle variations of peak pressure have been attributed to the lowsampling frequency used82 Both a direct measurement as well asthe integration of the pressure histories were used to estimate the av-erage Isp from hydrogenndashair systems to be between 1116 and 1333 sThe effect of the predetonator volume on the reduction of speci cimpulse was also shown to be a critical issue The effects of inletsand nozzles were not considered in this study

A conceptual design of a multitube PDE with a single air inletduct was discussed by Pegg et al83 Time-dependent computational

Fig 17 Pressure history at one location from a PDE operating at 5 Hzfor 20 cycles (from Ref 82)

uid dynamics analysis indicated that the inlet isolaterdiffusercon-cept would work and did not allow enough time for the formationof destabilizing hammer shocks Performance analysis includingcomponent ef ciencies reported in the paper indicate that opera-tional frequencies of the order of 75ndash100 Hz are required for theMach 12ndash3 ights considered Experimental data indicating mul-ticycle operation at 100 Hz with a rich ethylenendashoxygen mixturewere presented by Sterling et al84 DDT enhancement devices andpredetonators were used in the experiments In spite of this det-onation waves did not occur all of the time and when detonationwaves failed the tube became very hot and testing was suspendedTherefore the maximum testing time was reported to be 05 s Inspite of this limitation the two studies8384 taken together high-light the possibility of PDEs burning hydrocarbon fuels for realisticmissions

Five different nozzle shapes and their effect on performance werestudied computationally by Cambier and Tegner85 Their results in-dicate that the presence of a nozzle can affect the performance ofthe PDE by increasing the thrust delivery during the ignition phaseThe bell-shaped nozzles appeared to give higher performance thanshapes with positive curvature Their results also showed that noz-zles also affected the ow dynamics and hence the timing of thevarious phases of the engine cycle In an experimental study of aconical exhaust nozzle the authors of Ref 86 noted improved per-formance (based on higher velocities which were still sub-CndashJ) butdid not see any effect on the blowdown process More recently87

the effectof various nozzle shapes including converging divergingand straight have been reexamined computationally The converg-ing sections of nozzles introduced shock wave re ections whereasdiverging sections generated a negative thrust for a portion of thecycle due to overexpansion In spite of these limitations the overallconclusion of these studies was that ldquonozzles can drastically in-crease ef ciency of the PDEsrdquo87 There is also some experimentalevidence suggesting that propulsive performances can be increasedby the addition of a nozzle73 However factors such as the effect ofthe nozzles on detonation transmission and the detailed dynamicsof the ow have not yet been elucidated This appears to be an areathat needs further investigation

Other recent efforts have focused on demonstrating the opera-tion of single- and multitube combustors coupled to an inlet using arotary-valvemechanism88 The valve servesto both meter the air owinto the combustor and to isolate the inlet from the high pressuresproduced during the detonation cycle Utilizing multiple combus-tors that ll and detonate out of phase allows the continual use of thein owing air Firing rates of up to 12 Hz per combustor were demon-strated for a hydrogen-fueled system Several other issues related tothe practical development of PDEs have been discussed by Bussinget al89 According to them three of the more important engineer-ing issues are the development of a valve-injection subsystem anignition system and a low-volume predetonator

The use of PDEs for rockets has been revived recently90 Therocket mode of operation is very similar to the description providedearlier of the airbreathing engine with ignition at the closed endexcept that the oxidizer also needs to be injected into the system pe-riodically An advantage of PDEs for rockets would be their higherpower density thus enabling the development of more compact

1706 KAILASANATH

rockets Reference 90 shows the pressure traces from a pulsed det-onation rocket engine operating at 145 Hz on a hydrogenndashoxygenmixture

More recently a two-tube rotary-valved PDE with ight-sizecomponents was operated at 40 Hz per combustor for 30 s withan ethylenendashair mixture91 In addition throttling of the device al-lowed operation at three thrust levels during a 10-s demonstrationexperiment Contrary to some earlier estimates the thermal loadswere signi cant and it was not possible to operate for more than3 s without water cooling It is not clear if this was because ofperiodic detonation failures and consequent de agrative modes ofcombustion as reported in some earlier studies

The work reported in the open literature and discussed earlier hasfocused on gaseous fuels However for volume-limited propulsionapplications PDEs operating on liquid fuels need to be demon-strated In a recent study Brophy et al92 reported on experimentsusing JP-10oxygen and JP-10air aerosols The fuelndashoxygen mix-ture was successfully detonated to obtain an engine operating at5 Hz but the tests with the fuelndashair mixture were not successful

Computational studies continue to make progress in the study ofthe PDE8793 iexcl 97 The effect of nozzle shapes and incomplete transi-tion to CndashJ detonations have been studied by Eidelman and Yang87

They nd that the cycle ef ciency will be virtually the same whethera CndashJ detonation wave is initiated instantaneously or if the CndashJ valueis attainedonly at the end of the detonation chamber Primarilybasedon this study87 they come to an interesting conclusion that ldquoinitia-tion energies that are required for PDE operations can be small andcomparable to energies required for initiation of (other) combustionprocessesrdquo However this is yet to be substantiated experimentallySekar et al93 reexamine open- and closed-end initiation and thebasic self-aspirating con guration of Eidelman8 using more recentcomputational tools Injection and mixing issues that have gener-ally been ignored in previous numerical studies are also beginningto be addressed94 Both front-wall and lateral-wall injections werestudied and the time required for fuel injection was estimated Itis suggested that it may be possible to ll only a portion of thechamber during high-frequency operations Recent computationalstudies also suggest that preconditioning the fuelndashair mixture us-ing a shock wave can signi cantly reduce the DDT distance95 Theperformance estimates mentioned throughout the discussion of thePDE show signi cant variation A systematic review of various per-formance estimates and possible reasons for the observed variationhave been presented9697 These numerical studies suggest that thetime history of the back pressure is a crucial parameter affecting theestimated performance

Other Detonation EnginesThere have also been several other interesting conceptual studies

on the applications of detonations to engines Cambier et al98 haveproposed a pulsed-detonation wave augmentation device as part ofa hybrid engine for a single-stage-to-orbit airbreathing hypersonicvehicle Here the PDE is used for both thrust generation as well asfor mixingcombustion augmentation Two speci c con gurationswere investigated numerically

The concept of a detonation internal combustion engine was pro-posed by Loth and Loth99 and Loth100 Here a conventional pistoncylinder con guration is modi ed to incorporate a separate deto-nation chamber which is isolated by a valve from the compressionchamber during fuel injection and discharges tangentially into a vor-texchamber formedby the piston and cylinder at top deadcenterTherapid detonative combustion is followed by rapid dilution throughmixing with air in the vortex chamber to reduce the formation ofNOx and unburned hydrocarbons and to achieve an overall leanmixture The vortex chamber was also designed to store a portion ofthe detonation waversquos kinetic energy Overall thermal ef cienciessomewhat better than constant volume combustion were postulatedunder ideal conditions

ConclusionsDetonation waves have been explored extensively for propulsion

applications because of their inherent theoretical advantage overde agrative combustion However practical developments to date

have been of idealized systems or laboratory-scale devices or bothThe basic ideas behind most of the current systems being investi-gated have been known for a while but still many of the details needto be understood Advances in computations and experimental di-agnostics appear to be poised to make practical propulsion devicesbased on detonation waves a reality

Two major systems currently under development are the ram ac-celerator and the PDE In both cases the theoretical and computa-tional studies have been far more encouraging than the experimentsHowever in both cases actual demonstrations of devices workingon gaseous fuels have taken place In the case of the ram accelera-tor fuelndashoxygenndashdiluent mixtures have been used and the primaryproblem appears to be heat transfer and related material failuresThe use of gaseous fuels is not a major issue here but new materialsfor the projectiles need to be explored

In the case of the PDE most of the successful operations havebeen with fuelndashoxygen mixtures There still has not been a directcomparison between detailed experimental data and related theoret-ical and computational results Furthermore most of the PDE andrelated work to date have focused on gaseous mixtures PDEs op-erating on multiphase mixtures need to be emphasized because itwould be more practical to use liquid fuels for most of the propul-sion applicationsproposed for these systemsThere are several othergeneral issues that need to be resolved Clearly the evacuation ofexhaust gases is important both from the point of view of controllingthe cycle time and for preventing autoignition and mixing with freshgases The heat transfer and noise from PDEs have also not beenreported in any detail Other issues include inletndashcombustor andcombustorndashnozzle interactions and scalability of the entire device

AcknowledgmentsThis work has been sponsored by the Of ce of Naval Research

through the Mechanics and Energy Conversion Division and theUS Naval Research Laboratory

References1Fickett W and Davis W C Detonation Univ of California Press

Berkeley CA 19792Hoffmann N ldquoReaction Propulsion by Intermittent Detonative Com-

bustionrdquo German Ministry of Supply AI152365 Volkenrode Translation1940 (cited in Ref 16)

3Roy M ldquoPropulsion par Statoreacteur a Detonationrdquo Comptes-Rendusde lrsquoAcademie des Sciences Vol 222 1946 pp 31 32

4Dabora E K and Broda J C ldquoStanding Normal Detonations andOblique Detonations for Propulsionrdquo AIAA Paper 93-2325 July 1993

5Dabora E K ldquoStatus of Gaseous Detonation Waves and Their Rolein Propulsionrdquo Fall Technical Meeting of the Eastern States Section ofthe Combustion Institute Combustion Inst Pittsburgh PA 1994 pp 11ndash

186Rubins P M and Bauer R C ldquoReview of Shock-Induced Supersonic

Combustion Research and Hypersonic Applicationsrdquo Journal of Propulsionand Power Vol 10 No 5 1994 pp 593ndash601

7Eidelman S Grossmann W and Lottati I ldquoReview of PropulsionApplications and Numerical Simulations of the Pulsed Detonation EngineConceptrdquo Journal of Propulsion and Power Vol 7 No 6 1991 pp 857ndash

865 also AIAA Paper 89-2446 July 19898Eidelman S and Grossmann W ldquoPulsed Detonation Engine Experi-

mental and Theoretical Reviewrdquo AIAA Paper 92-3168 July 19929Reingold L ldquoRecherches sur les combustions permanentes apportees

aux foyers a circulation interne supersoniquerdquo ONERA TN2 Dept of En-ergy and Propulsion Study 728-E Paris March 1950

10Bitondo D and Bollay W ldquoPreliminary Performance Analysis ofthe Pulse-Detonation-Jet Engine Systemrdquo Aerophysics Development CorpRept ADC-102-1 April 1952

11Gross R A ldquoExploratory Study of Combustion in Supersonic FlowrdquoUS Air Force Of ce of Scienti c Research TN 59-587 Washington DCJune 1959

12Gross R A and Chinitz W ldquoA Study of Supersonic CombustionrdquoJournal of Aerospace Science Vol 27 No 7 1960 pp 517ndash525

13Nicholls J A Dabora E K and Gealer R L ldquoStudies in Con-nection with Stabilized Gaseous Detonation Wavesrdquo Seventh Symposium(International) on Combustion Combustion Inst Pittsburgh PA 1959 pp766ndash772

14Nicholls J A and Dabora E K ldquoRecent Results on Standing Detona-tion Wavesrdquo Eighth Symposium (International) on Combustion CombustionInst Pittsburgh PA 1962 pp 644ndash655

KAILASANATH 1707

15Dunlap R Brehm R L and Nicholls J A ldquoA Preliminary Studyof the Application of Steady-State Detonative Combustion to a ReactionEnginerdquo Jet Propulsion Vol 28 No 7 1958 pp 451ndash456

16Nicholls J A Wilkinson H R and Morrison R B ldquoIntermittentDetonation as a Thrust-Producing Mechanismrdquo Jet Propulsion Vol 27No 5 1957 pp 534ndash541

17Krzycki L J ldquoPerformance Characteristics of an Intermittent Detona-tion Devicerdquo US Naval Ordnance Test Station US Naval Weapons Rept7655 China Lake CA 1962

18Nicholls J A Cullen R E and Ragland K W ldquoFeasibility Studiesof a Rotating Detonation Wave Rocket Motorrdquo Journal of Spacecraft andRockets Vol 3 No 6 1966 pp 893ndash898

19Adamson T C and Olsson G R ldquoPerformance Analysis of a RotatingDetonation Wave Rocket Enginerdquo Astronautica Acta Vol 13 No 4 1967pp 405ndash415

20Shen P I and Adamson T C ldquoTheoretical Analysis of a RotatingTwo-Phase Detonation in Liquid Rocket Motorsrdquo Astronautica Acta Vol17 No 4 1972 pp 715ndash728

21Clayton R M and Rogero R S ldquoExperimental Measurements on aRotating Detonation-Like Wave Observed During Liquid Rocket ResonantCombustionrdquo Jet Propulsion Lab TR 32-788 California Inst of Technol-ogy Pasadena CA Aug 1965

22Back L H ldquoApplication of Blast Wave Theory to Explosive Propul-sionrdquo Acta Astronautica Vol 2 1975 pp 391ndash407

23Varsi G Back L H and Kim K ldquoBlast Wave in a Nozzle for Propul-sion Applicationsrdquo Acta Astronautica Vol 3 1976 pp 141ndash156

24Kim K Varsi G and Back LH ldquoBlast Wave Analysis for DetonationPropulsionrdquo AIAA Journal Vol 10 No 10 1977 pp 1500ndash1502

25Back L H Dowler W L and Varsi G ldquoDetonation PropulsionExperiments and Theoryrdquo AIAA Journal Vol 21 No 10 1983 pp 1418ndash

142726Kantrowitz A ldquoPropulsion to Orbit by Ground Based Lasersrdquo Astro-

nautics and Aeronautics Vol 10 No 5 1972 p 7427Hyde R A ldquoOne-D Modelling of a Two-Pulse LSD Thrusterrdquo Pro-

ceedings of the 1986 SDIODARPA Workshop on Laser Propulsion editedby J T Kare Lawrence Livermore National Labs 1987 pp 79ndash88

28Emery M H Kailasanath K Oran E S and Gardner J H ldquoCompu-tational Studies of Laser Supported Detonationsrdquo Proceedings of the 1987SDIO Workshop on Laser Propulsion edited by J T Kare Lawrence Liv-ermore National Labs 1990 pp 43ndash55

29Emery M H Kailasanath K and Oran E S ldquoNumerical Simu-lations of Laser Supported Detonationsrdquo 1988 SDIO Workshop on LaserPropulsion Lawrence Livermore National Labs April 1988

30Carrier G F Fendell F McGregor D Cook S and Vazirani MldquoLaser-Initiated Conical Detonation Wave for Supersonic CombustionrdquoAIAA Paper 91-0578 Jan 1991

31Fendell F E Mitchell J McGregor R Magiawala K and Shef eldM ldquoLaser-Initiated Conical Detonation Wave for Supersonic CombustionIIrdquo AIAA Paper 92-0088 Jan 1992

32Carrier G F Fendell F E and Chou M S ldquoLaser-Initiated ConicalDetonation Wave for Supersonic Combustion IIIrdquo AIAA Paper 92-3247July 1992

33Menees G P Adelman H G Cambier J-L and Bowles J V ldquoWaveCombustors for Trans-Atmospheric Vehiclesrdquo Journal of Propulsion andPower Vol 8 No 3 1992 pp 709ndash713

34Cambier J L Adelman H and Menees G ldquoNumerical Simulationsof an Oblique Detonation Wave Enginerdquo AIAA Paper 88-0063 1988

35Atamanchuk T and Sislian J ldquoOn- and Off-design Performance Anal-ysis of Hypersonic Detonation Wave Ramjetsrdquo AIAA Paper 90-2473 1990

36Kuznetsov M M Neyland V J and Sayapin G N ldquoEf ciency In-vestigation of Scramjet with Detonation and Shockless Supersonic Combus-tionrdquo Utchenye Zapisky TsAGI Vol 23 No 2 1992 pp 30ndash37

37Bruckner A P and Hertzberg A ldquoRam Accelerator Direct LaunchSystem for Space Cargordquo International Astronautical Federation Paper 87-211 1987

38Hertzberg A Bruckner A P and Bogdanoff D W ldquoRam Accel-erator A New Chemical Method for Accelerating Projectiles to UltrahighVelocitiesrdquo AIAA Journal Vol 26 No 2 1988 pp 195ndash203

39Kull A Burnham E Knowlen C Bruckner A P and HertzbergA ldquoExperimental Studies of the Superdetonative Ram Accelerator ModesrdquoAIAA Paper 89-2632 1989

40Patz G Seiler F Smeets G and Srulijes J ldquoStatus of ISLrsquos RA-MAC30 with Fin Guided Projectiles Accelerated in a Smooth Borerdquo Pro-ceedings of the Second International Workshop on Ram Accelerators Univof Washington Seattle WA 1995

41Seiler F Patz G Smeets G and Srulijes J ldquoGasdynamic Limitsof Ignition and Combustion of a Gas Mixture in ISLrsquos RAMAC30 ScramAcceleratorrdquo Proceedings of the Second International Workshop on RamAccelerators Univ of Washington Seattle WA 1995

42Seiler F Patz G Smeets G and Srulijes J ldquoIn uence of ProjectileMaterial and Gas Composition on Superdetonative Combustion in ISLrsquosRAMAC30rdquo AIAA Paper 98-3445 July 1998

43Brackett D C and Bogdanoff D W ldquoComputational Investigation ofOblique Detonation Ramjet-in-Tube Conceptsrdquo Journal of Propulsion andPower Vol 5 No 3 1989 pp 276ndash281

44Yungster S Eberhardt S and Bruckner A P ldquoNumerical Simula-tion of Shock-Induced Combustion Generated by High-Speed Projectiles inDetonable Gas Mixturesrdquo AIAA Paper 89-0673 1989

45Rom J and Kivity Y ldquoAccelerating Projectiles Up to 12 kms Utiliz-ing the Continuous Detonation Propulsion Methodrdquo AIAA Paper 88-29691988

46Humphrey J W ldquoParametric Study of an ODW Scramaccelerator forHypersonic Test Facilitiesrdquo AIAA Paper 90-2470 1990

47Yungster S ldquoNavierndashStokes Simulation of the Supersonic CombustionFlow eld in a Ram Acceleratorrdquo AIAA Paper 91-1916 1991

48Li C Kailasanath K Oran E S Boris J P Paper and LandsbergA M ldquoNumerical Simulations of Transient Flows in Ram AcceleratorsrdquoAIAA Paper 93-1916 1993

49Li C Kailasanath K Oran E S Landsberg A M and Boris J PldquoDynamics of Oblique Detonations in Ram Acceleratorsrdquo Shock Waves Vol5 June 1995 pp 97ndash101

50Yungster S and Radhakrishnan K ldquoComputational Study of FlowEstablishment in a Ram Acceleratorrdquo AIAA Paper 98-2489 July 1995

51Choi J Y Jeung I-S and Yoon Y ldquoNumerical Study of ScramAccelerator Starting Characteristicsrdquo AIAA Journal Vol 36 No 6 1998pp 1029ndash1038

52Bezgin L Ganzhelo A Gouskov O and Kopchenov V ldquoSome Nu-merical Investigation Results of Shock-Induced Combustionrdquo Proceedingsof the AIAA 8th International Space Planes and Hypersonic Systems andTechnologies Conference AIAA Reston VA 1998 pp 72ndash82

53Pratt D W Humphrey J W and Glenn D E ldquoMorphology of Stand-ing Oblique Detonation Wavesrdquo Journal of Propulsion and Power Vol 7No 5 1991

54Fujiwara T Matsuo A and Nomoto H ldquoTwo-Dimensional Detona-tion Supported by a Blunt Body or a Wedgerdquo AIAA Paper 88-0089 Jan1988

55Fan B C Sichel M and Kauffman C W ldquoAnalysis of ObliqueShock-Detonation Wave Interactions in the Supersonic Flow of a Com-bustible Mediumrdquo AIAA Paper 88-0441 Jan 1988

56Dabora E K Desbordes D Guerraud C and Wagner H GldquoOblique Detonations at Hypersonic Velocitiesrdquo Dynamics of Detonationsand Explosions Detonations Vol 133 Progress in Astronautics and Aero-nautics AIAA Washington DC 1990 pp 187ndash201

57Li C Kailasanath K and Oran E S ldquoStability of Oblique Detona-tions in Ram Acceleratorsrdquo AIAA Paper 92-0089 Jan 1992

58Li C Kailasanath K and Oran E S ldquoStructure of Reaction WavesBehind Oblique Shocksrdquo Dynamic Aspects of Detonation Vol 153 Progressin Astronautics and Aeronautics AIAA Washington DC 1993 pp 231ndash

24059Li C Kailasanath K and Oran E S ldquoDetonation Structures Behind

Oblique Shocksrdquo Physics of Fluids Vol 6 April 1994 pp 1600ndash161160Da Silva L F and Deshaies B ldquoNumerical Study of Some Ignition

Regimes of Combustible Supersonic Flows over a Wedgerdquo Proceedings ofSecond International Workshop on Ram Accelerators Paper 15 Univ ofWashington Seattle WA 1995

61Thaker A A and Chelliah H K ldquoNumerical Prediction of ObliqueDetonation Wave Structures Using Detailed and Reduced Reaction Mecha-nismsrdquo Combustion Theory and Modelling Vol 1 No 4 1997 pp 347ndash376

62Desbordes D Guerraud C and Hamada L ldquoSupersonic H2-AirCombustions Behind Oblique Shock Wavesrdquo 14th International Colloquiumon Dynamics of Explosions and Reactive Systems Paper D28 Univ ofCoimbra Portugal Aug 1993

63Li C Kailasanath K and Oran E S ldquoDetonation Structures Gener-ated by Multiple Shocks on Ram-Accelerator Projectilesrdquo Combustion andFlame Vol 108 1997 pp 173ndash186

64Viguier C and Desbordes D ldquoConditions of Onset and Stabilizationof Oblique Detonation Wavesrdquo Advances in Experimentation and Compu-tation of Detonations edited by G D Roy S H Frolov K Kailasanathand N Smirnov ENAS Moscow 1998 pp 40 41

65Bussing Tand Pappas G ldquoIntroduction toPulseDetonation EnginesrdquoAIAA Paper 94-0263 1994

66Helman D Shreeve R P and Eidelman S ldquoDetonation Pulse En-ginerdquo AIAA Paper 86-1683 June 1986

67Cambier J L and Adelman H G ldquoPreliminary Numerical Simu-lations of a Pulsed Detonation Wave Enginerdquo AIAA Paper 88-2960 July1988

68Eidelman S Grossmann W and Lottati I ldquoAir-Breathing PulsedDetonationEngine Concept A Numerical Studyrdquo AIAA Paper 90-2420 July1990

69Eidelman S Grossmann W and Lottati I ldquoComputational Analysisof Pulsed Detonation Engines and Applicationsrdquo AIAA Paper 90-0460 Jan1990

1706 KAILASANATH

rockets Reference 90 shows the pressure traces from a pulsed det-onation rocket engine operating at 145 Hz on a hydrogenndashoxygenmixture

More recently a two-tube rotary-valved PDE with ight-sizecomponents was operated at 40 Hz per combustor for 30 s withan ethylenendashair mixture91 In addition throttling of the device al-lowed operation at three thrust levels during a 10-s demonstrationexperiment Contrary to some earlier estimates the thermal loadswere signi cant and it was not possible to operate for more than3 s without water cooling It is not clear if this was because ofperiodic detonation failures and consequent de agrative modes ofcombustion as reported in some earlier studies

The work reported in the open literature and discussed earlier hasfocused on gaseous fuels However for volume-limited propulsionapplications PDEs operating on liquid fuels need to be demon-strated In a recent study Brophy et al92 reported on experimentsusing JP-10oxygen and JP-10air aerosols The fuelndashoxygen mix-ture was successfully detonated to obtain an engine operating at5 Hz but the tests with the fuelndashair mixture were not successful

Computational studies continue to make progress in the study ofthe PDE8793 iexcl 97 The effect of nozzle shapes and incomplete transi-tion to CndashJ detonations have been studied by Eidelman and Yang87

They nd that the cycle ef ciency will be virtually the same whethera CndashJ detonation wave is initiated instantaneously or if the CndashJ valueis attainedonly at the end of the detonation chamber Primarilybasedon this study87 they come to an interesting conclusion that ldquoinitia-tion energies that are required for PDE operations can be small andcomparable to energies required for initiation of (other) combustionprocessesrdquo However this is yet to be substantiated experimentallySekar et al93 reexamine open- and closed-end initiation and thebasic self-aspirating con guration of Eidelman8 using more recentcomputational tools Injection and mixing issues that have gener-ally been ignored in previous numerical studies are also beginningto be addressed94 Both front-wall and lateral-wall injections werestudied and the time required for fuel injection was estimated Itis suggested that it may be possible to ll only a portion of thechamber during high-frequency operations Recent computationalstudies also suggest that preconditioning the fuelndashair mixture us-ing a shock wave can signi cantly reduce the DDT distance95 Theperformance estimates mentioned throughout the discussion of thePDE show signi cant variation A systematic review of various per-formance estimates and possible reasons for the observed variationhave been presented9697 These numerical studies suggest that thetime history of the back pressure is a crucial parameter affecting theestimated performance

Other Detonation EnginesThere have also been several other interesting conceptual studies

on the applications of detonations to engines Cambier et al98 haveproposed a pulsed-detonation wave augmentation device as part ofa hybrid engine for a single-stage-to-orbit airbreathing hypersonicvehicle Here the PDE is used for both thrust generation as well asfor mixingcombustion augmentation Two speci c con gurationswere investigated numerically

The concept of a detonation internal combustion engine was pro-posed by Loth and Loth99 and Loth100 Here a conventional pistoncylinder con guration is modi ed to incorporate a separate deto-nation chamber which is isolated by a valve from the compressionchamber during fuel injection and discharges tangentially into a vor-texchamber formedby the piston and cylinder at top deadcenterTherapid detonative combustion is followed by rapid dilution throughmixing with air in the vortex chamber to reduce the formation ofNOx and unburned hydrocarbons and to achieve an overall leanmixture The vortex chamber was also designed to store a portion ofthe detonation waversquos kinetic energy Overall thermal ef cienciessomewhat better than constant volume combustion were postulatedunder ideal conditions

ConclusionsDetonation waves have been explored extensively for propulsion

applications because of their inherent theoretical advantage overde agrative combustion However practical developments to date

have been of idealized systems or laboratory-scale devices or bothThe basic ideas behind most of the current systems being investi-gated have been known for a while but still many of the details needto be understood Advances in computations and experimental di-agnostics appear to be poised to make practical propulsion devicesbased on detonation waves a reality

Two major systems currently under development are the ram ac-celerator and the PDE In both cases the theoretical and computa-tional studies have been far more encouraging than the experimentsHowever in both cases actual demonstrations of devices workingon gaseous fuels have taken place In the case of the ram accelera-tor fuelndashoxygenndashdiluent mixtures have been used and the primaryproblem appears to be heat transfer and related material failuresThe use of gaseous fuels is not a major issue here but new materialsfor the projectiles need to be explored

In the case of the PDE most of the successful operations havebeen with fuelndashoxygen mixtures There still has not been a directcomparison between detailed experimental data and related theoret-ical and computational results Furthermore most of the PDE andrelated work to date have focused on gaseous mixtures PDEs op-erating on multiphase mixtures need to be emphasized because itwould be more practical to use liquid fuels for most of the propul-sion applicationsproposed for these systemsThere are several othergeneral issues that need to be resolved Clearly the evacuation ofexhaust gases is important both from the point of view of controllingthe cycle time and for preventing autoignition and mixing with freshgases The heat transfer and noise from PDEs have also not beenreported in any detail Other issues include inletndashcombustor andcombustorndashnozzle interactions and scalability of the entire device

AcknowledgmentsThis work has been sponsored by the Of ce of Naval Research

through the Mechanics and Energy Conversion Division and theUS Naval Research Laboratory

References1Fickett W and Davis W C Detonation Univ of California Press

Berkeley CA 19792Hoffmann N ldquoReaction Propulsion by Intermittent Detonative Com-

bustionrdquo German Ministry of Supply AI152365 Volkenrode Translation1940 (cited in Ref 16)

3Roy M ldquoPropulsion par Statoreacteur a Detonationrdquo Comptes-Rendusde lrsquoAcademie des Sciences Vol 222 1946 pp 31 32

4Dabora E K and Broda J C ldquoStanding Normal Detonations andOblique Detonations for Propulsionrdquo AIAA Paper 93-2325 July 1993

5Dabora E K ldquoStatus of Gaseous Detonation Waves and Their Rolein Propulsionrdquo Fall Technical Meeting of the Eastern States Section ofthe Combustion Institute Combustion Inst Pittsburgh PA 1994 pp 11ndash

186Rubins P M and Bauer R C ldquoReview of Shock-Induced Supersonic

Combustion Research and Hypersonic Applicationsrdquo Journal of Propulsionand Power Vol 10 No 5 1994 pp 593ndash601

7Eidelman S Grossmann W and Lottati I ldquoReview of PropulsionApplications and Numerical Simulations of the Pulsed Detonation EngineConceptrdquo Journal of Propulsion and Power Vol 7 No 6 1991 pp 857ndash

865 also AIAA Paper 89-2446 July 19898Eidelman S and Grossmann W ldquoPulsed Detonation Engine Experi-

mental and Theoretical Reviewrdquo AIAA Paper 92-3168 July 19929Reingold L ldquoRecherches sur les combustions permanentes apportees

aux foyers a circulation interne supersoniquerdquo ONERA TN2 Dept of En-ergy and Propulsion Study 728-E Paris March 1950

10Bitondo D and Bollay W ldquoPreliminary Performance Analysis ofthe Pulse-Detonation-Jet Engine Systemrdquo Aerophysics Development CorpRept ADC-102-1 April 1952

11Gross R A ldquoExploratory Study of Combustion in Supersonic FlowrdquoUS Air Force Of ce of Scienti c Research TN 59-587 Washington DCJune 1959

12Gross R A and Chinitz W ldquoA Study of Supersonic CombustionrdquoJournal of Aerospace Science Vol 27 No 7 1960 pp 517ndash525

13Nicholls J A Dabora E K and Gealer R L ldquoStudies in Con-nection with Stabilized Gaseous Detonation Wavesrdquo Seventh Symposium(International) on Combustion Combustion Inst Pittsburgh PA 1959 pp766ndash772

14Nicholls J A and Dabora E K ldquoRecent Results on Standing Detona-tion Wavesrdquo Eighth Symposium (International) on Combustion CombustionInst Pittsburgh PA 1962 pp 644ndash655

KAILASANATH 1707

15Dunlap R Brehm R L and Nicholls J A ldquoA Preliminary Studyof the Application of Steady-State Detonative Combustion to a ReactionEnginerdquo Jet Propulsion Vol 28 No 7 1958 pp 451ndash456

16Nicholls J A Wilkinson H R and Morrison R B ldquoIntermittentDetonation as a Thrust-Producing Mechanismrdquo Jet Propulsion Vol 27No 5 1957 pp 534ndash541

17Krzycki L J ldquoPerformance Characteristics of an Intermittent Detona-tion Devicerdquo US Naval Ordnance Test Station US Naval Weapons Rept7655 China Lake CA 1962

18Nicholls J A Cullen R E and Ragland K W ldquoFeasibility Studiesof a Rotating Detonation Wave Rocket Motorrdquo Journal of Spacecraft andRockets Vol 3 No 6 1966 pp 893ndash898

19Adamson T C and Olsson G R ldquoPerformance Analysis of a RotatingDetonation Wave Rocket Enginerdquo Astronautica Acta Vol 13 No 4 1967pp 405ndash415

20Shen P I and Adamson T C ldquoTheoretical Analysis of a RotatingTwo-Phase Detonation in Liquid Rocket Motorsrdquo Astronautica Acta Vol17 No 4 1972 pp 715ndash728

21Clayton R M and Rogero R S ldquoExperimental Measurements on aRotating Detonation-Like Wave Observed During Liquid Rocket ResonantCombustionrdquo Jet Propulsion Lab TR 32-788 California Inst of Technol-ogy Pasadena CA Aug 1965

22Back L H ldquoApplication of Blast Wave Theory to Explosive Propul-sionrdquo Acta Astronautica Vol 2 1975 pp 391ndash407

23Varsi G Back L H and Kim K ldquoBlast Wave in a Nozzle for Propul-sion Applicationsrdquo Acta Astronautica Vol 3 1976 pp 141ndash156

24Kim K Varsi G and Back LH ldquoBlast Wave Analysis for DetonationPropulsionrdquo AIAA Journal Vol 10 No 10 1977 pp 1500ndash1502

25Back L H Dowler W L and Varsi G ldquoDetonation PropulsionExperiments and Theoryrdquo AIAA Journal Vol 21 No 10 1983 pp 1418ndash

142726Kantrowitz A ldquoPropulsion to Orbit by Ground Based Lasersrdquo Astro-

nautics and Aeronautics Vol 10 No 5 1972 p 7427Hyde R A ldquoOne-D Modelling of a Two-Pulse LSD Thrusterrdquo Pro-

ceedings of the 1986 SDIODARPA Workshop on Laser Propulsion editedby J T Kare Lawrence Livermore National Labs 1987 pp 79ndash88

28Emery M H Kailasanath K Oran E S and Gardner J H ldquoCompu-tational Studies of Laser Supported Detonationsrdquo Proceedings of the 1987SDIO Workshop on Laser Propulsion edited by J T Kare Lawrence Liv-ermore National Labs 1990 pp 43ndash55

29Emery M H Kailasanath K and Oran E S ldquoNumerical Simu-lations of Laser Supported Detonationsrdquo 1988 SDIO Workshop on LaserPropulsion Lawrence Livermore National Labs April 1988

30Carrier G F Fendell F McGregor D Cook S and Vazirani MldquoLaser-Initiated Conical Detonation Wave for Supersonic CombustionrdquoAIAA Paper 91-0578 Jan 1991

31Fendell F E Mitchell J McGregor R Magiawala K and Shef eldM ldquoLaser-Initiated Conical Detonation Wave for Supersonic CombustionIIrdquo AIAA Paper 92-0088 Jan 1992

32Carrier G F Fendell F E and Chou M S ldquoLaser-Initiated ConicalDetonation Wave for Supersonic Combustion IIIrdquo AIAA Paper 92-3247July 1992

33Menees G P Adelman H G Cambier J-L and Bowles J V ldquoWaveCombustors for Trans-Atmospheric Vehiclesrdquo Journal of Propulsion andPower Vol 8 No 3 1992 pp 709ndash713

34Cambier J L Adelman H and Menees G ldquoNumerical Simulationsof an Oblique Detonation Wave Enginerdquo AIAA Paper 88-0063 1988

35Atamanchuk T and Sislian J ldquoOn- and Off-design Performance Anal-ysis of Hypersonic Detonation Wave Ramjetsrdquo AIAA Paper 90-2473 1990

36Kuznetsov M M Neyland V J and Sayapin G N ldquoEf ciency In-vestigation of Scramjet with Detonation and Shockless Supersonic Combus-tionrdquo Utchenye Zapisky TsAGI Vol 23 No 2 1992 pp 30ndash37

37Bruckner A P and Hertzberg A ldquoRam Accelerator Direct LaunchSystem for Space Cargordquo International Astronautical Federation Paper 87-211 1987

38Hertzberg A Bruckner A P and Bogdanoff D W ldquoRam Accel-erator A New Chemical Method for Accelerating Projectiles to UltrahighVelocitiesrdquo AIAA Journal Vol 26 No 2 1988 pp 195ndash203

39Kull A Burnham E Knowlen C Bruckner A P and HertzbergA ldquoExperimental Studies of the Superdetonative Ram Accelerator ModesrdquoAIAA Paper 89-2632 1989

40Patz G Seiler F Smeets G and Srulijes J ldquoStatus of ISLrsquos RA-MAC30 with Fin Guided Projectiles Accelerated in a Smooth Borerdquo Pro-ceedings of the Second International Workshop on Ram Accelerators Univof Washington Seattle WA 1995

41Seiler F Patz G Smeets G and Srulijes J ldquoGasdynamic Limitsof Ignition and Combustion of a Gas Mixture in ISLrsquos RAMAC30 ScramAcceleratorrdquo Proceedings of the Second International Workshop on RamAccelerators Univ of Washington Seattle WA 1995

42Seiler F Patz G Smeets G and Srulijes J ldquoIn uence of ProjectileMaterial and Gas Composition on Superdetonative Combustion in ISLrsquosRAMAC30rdquo AIAA Paper 98-3445 July 1998

43Brackett D C and Bogdanoff D W ldquoComputational Investigation ofOblique Detonation Ramjet-in-Tube Conceptsrdquo Journal of Propulsion andPower Vol 5 No 3 1989 pp 276ndash281

44Yungster S Eberhardt S and Bruckner A P ldquoNumerical Simula-tion of Shock-Induced Combustion Generated by High-Speed Projectiles inDetonable Gas Mixturesrdquo AIAA Paper 89-0673 1989

45Rom J and Kivity Y ldquoAccelerating Projectiles Up to 12 kms Utiliz-ing the Continuous Detonation Propulsion Methodrdquo AIAA Paper 88-29691988

46Humphrey J W ldquoParametric Study of an ODW Scramaccelerator forHypersonic Test Facilitiesrdquo AIAA Paper 90-2470 1990

47Yungster S ldquoNavierndashStokes Simulation of the Supersonic CombustionFlow eld in a Ram Acceleratorrdquo AIAA Paper 91-1916 1991

48Li C Kailasanath K Oran E S Boris J P Paper and LandsbergA M ldquoNumerical Simulations of Transient Flows in Ram AcceleratorsrdquoAIAA Paper 93-1916 1993

49Li C Kailasanath K Oran E S Landsberg A M and Boris J PldquoDynamics of Oblique Detonations in Ram Acceleratorsrdquo Shock Waves Vol5 June 1995 pp 97ndash101

50Yungster S and Radhakrishnan K ldquoComputational Study of FlowEstablishment in a Ram Acceleratorrdquo AIAA Paper 98-2489 July 1995

51Choi J Y Jeung I-S and Yoon Y ldquoNumerical Study of ScramAccelerator Starting Characteristicsrdquo AIAA Journal Vol 36 No 6 1998pp 1029ndash1038

52Bezgin L Ganzhelo A Gouskov O and Kopchenov V ldquoSome Nu-merical Investigation Results of Shock-Induced Combustionrdquo Proceedingsof the AIAA 8th International Space Planes and Hypersonic Systems andTechnologies Conference AIAA Reston VA 1998 pp 72ndash82

53Pratt D W Humphrey J W and Glenn D E ldquoMorphology of Stand-ing Oblique Detonation Wavesrdquo Journal of Propulsion and Power Vol 7No 5 1991

54Fujiwara T Matsuo A and Nomoto H ldquoTwo-Dimensional Detona-tion Supported by a Blunt Body or a Wedgerdquo AIAA Paper 88-0089 Jan1988

55Fan B C Sichel M and Kauffman C W ldquoAnalysis of ObliqueShock-Detonation Wave Interactions in the Supersonic Flow of a Com-bustible Mediumrdquo AIAA Paper 88-0441 Jan 1988

56Dabora E K Desbordes D Guerraud C and Wagner H GldquoOblique Detonations at Hypersonic Velocitiesrdquo Dynamics of Detonationsand Explosions Detonations Vol 133 Progress in Astronautics and Aero-nautics AIAA Washington DC 1990 pp 187ndash201

57Li C Kailasanath K and Oran E S ldquoStability of Oblique Detona-tions in Ram Acceleratorsrdquo AIAA Paper 92-0089 Jan 1992

58Li C Kailasanath K and Oran E S ldquoStructure of Reaction WavesBehind Oblique Shocksrdquo Dynamic Aspects of Detonation Vol 153 Progressin Astronautics and Aeronautics AIAA Washington DC 1993 pp 231ndash

24059Li C Kailasanath K and Oran E S ldquoDetonation Structures Behind

Oblique Shocksrdquo Physics of Fluids Vol 6 April 1994 pp 1600ndash161160Da Silva L F and Deshaies B ldquoNumerical Study of Some Ignition

Regimes of Combustible Supersonic Flows over a Wedgerdquo Proceedings ofSecond International Workshop on Ram Accelerators Paper 15 Univ ofWashington Seattle WA 1995

61Thaker A A and Chelliah H K ldquoNumerical Prediction of ObliqueDetonation Wave Structures Using Detailed and Reduced Reaction Mecha-nismsrdquo Combustion Theory and Modelling Vol 1 No 4 1997 pp 347ndash376

62Desbordes D Guerraud C and Hamada L ldquoSupersonic H2-AirCombustions Behind Oblique Shock Wavesrdquo 14th International Colloquiumon Dynamics of Explosions and Reactive Systems Paper D28 Univ ofCoimbra Portugal Aug 1993

63Li C Kailasanath K and Oran E S ldquoDetonation Structures Gener-ated by Multiple Shocks on Ram-Accelerator Projectilesrdquo Combustion andFlame Vol 108 1997 pp 173ndash186

64Viguier C and Desbordes D ldquoConditions of Onset and Stabilizationof Oblique Detonation Wavesrdquo Advances in Experimentation and Compu-tation of Detonations edited by G D Roy S H Frolov K Kailasanathand N Smirnov ENAS Moscow 1998 pp 40 41

65Bussing Tand Pappas G ldquoIntroduction toPulseDetonation EnginesrdquoAIAA Paper 94-0263 1994

66Helman D Shreeve R P and Eidelman S ldquoDetonation Pulse En-ginerdquo AIAA Paper 86-1683 June 1986

67Cambier J L and Adelman H G ldquoPreliminary Numerical Simu-lations of a Pulsed Detonation Wave Enginerdquo AIAA Paper 88-2960 July1988

68Eidelman S Grossmann W and Lottati I ldquoAir-Breathing PulsedDetonationEngine Concept A Numerical Studyrdquo AIAA Paper 90-2420 July1990

69Eidelman S Grossmann W and Lottati I ldquoComputational Analysisof Pulsed Detonation Engines and Applicationsrdquo AIAA Paper 90-0460 Jan1990

KAILASANATH 1707

15Dunlap R Brehm R L and Nicholls J A ldquoA Preliminary Studyof the Application of Steady-State Detonative Combustion to a ReactionEnginerdquo Jet Propulsion Vol 28 No 7 1958 pp 451ndash456

16Nicholls J A Wilkinson H R and Morrison R B ldquoIntermittentDetonation as a Thrust-Producing Mechanismrdquo Jet Propulsion Vol 27No 5 1957 pp 534ndash541

17Krzycki L J ldquoPerformance Characteristics of an Intermittent Detona-tion Devicerdquo US Naval Ordnance Test Station US Naval Weapons Rept7655 China Lake CA 1962

18Nicholls J A Cullen R E and Ragland K W ldquoFeasibility Studiesof a Rotating Detonation Wave Rocket Motorrdquo Journal of Spacecraft andRockets Vol 3 No 6 1966 pp 893ndash898

19Adamson T C and Olsson G R ldquoPerformance Analysis of a RotatingDetonation Wave Rocket Enginerdquo Astronautica Acta Vol 13 No 4 1967pp 405ndash415

20Shen P I and Adamson T C ldquoTheoretical Analysis of a RotatingTwo-Phase Detonation in Liquid Rocket Motorsrdquo Astronautica Acta Vol17 No 4 1972 pp 715ndash728

21Clayton R M and Rogero R S ldquoExperimental Measurements on aRotating Detonation-Like Wave Observed During Liquid Rocket ResonantCombustionrdquo Jet Propulsion Lab TR 32-788 California Inst of Technol-ogy Pasadena CA Aug 1965

22Back L H ldquoApplication of Blast Wave Theory to Explosive Propul-sionrdquo Acta Astronautica Vol 2 1975 pp 391ndash407

23Varsi G Back L H and Kim K ldquoBlast Wave in a Nozzle for Propul-sion Applicationsrdquo Acta Astronautica Vol 3 1976 pp 141ndash156

24Kim K Varsi G and Back LH ldquoBlast Wave Analysis for DetonationPropulsionrdquo AIAA Journal Vol 10 No 10 1977 pp 1500ndash1502

25Back L H Dowler W L and Varsi G ldquoDetonation PropulsionExperiments and Theoryrdquo AIAA Journal Vol 21 No 10 1983 pp 1418ndash

142726Kantrowitz A ldquoPropulsion to Orbit by Ground Based Lasersrdquo Astro-

nautics and Aeronautics Vol 10 No 5 1972 p 7427Hyde R A ldquoOne-D Modelling of a Two-Pulse LSD Thrusterrdquo Pro-

ceedings of the 1986 SDIODARPA Workshop on Laser Propulsion editedby J T Kare Lawrence Livermore National Labs 1987 pp 79ndash88

28Emery M H Kailasanath K Oran E S and Gardner J H ldquoCompu-tational Studies of Laser Supported Detonationsrdquo Proceedings of the 1987SDIO Workshop on Laser Propulsion edited by J T Kare Lawrence Liv-ermore National Labs 1990 pp 43ndash55

29Emery M H Kailasanath K and Oran E S ldquoNumerical Simu-lations of Laser Supported Detonationsrdquo 1988 SDIO Workshop on LaserPropulsion Lawrence Livermore National Labs April 1988

30Carrier G F Fendell F McGregor D Cook S and Vazirani MldquoLaser-Initiated Conical Detonation Wave for Supersonic CombustionrdquoAIAA Paper 91-0578 Jan 1991

31Fendell F E Mitchell J McGregor R Magiawala K and Shef eldM ldquoLaser-Initiated Conical Detonation Wave for Supersonic CombustionIIrdquo AIAA Paper 92-0088 Jan 1992

32Carrier G F Fendell F E and Chou M S ldquoLaser-Initiated ConicalDetonation Wave for Supersonic Combustion IIIrdquo AIAA Paper 92-3247July 1992

33Menees G P Adelman H G Cambier J-L and Bowles J V ldquoWaveCombustors for Trans-Atmospheric Vehiclesrdquo Journal of Propulsion andPower Vol 8 No 3 1992 pp 709ndash713

34Cambier J L Adelman H and Menees G ldquoNumerical Simulationsof an Oblique Detonation Wave Enginerdquo AIAA Paper 88-0063 1988

35Atamanchuk T and Sislian J ldquoOn- and Off-design Performance Anal-ysis of Hypersonic Detonation Wave Ramjetsrdquo AIAA Paper 90-2473 1990

36Kuznetsov M M Neyland V J and Sayapin G N ldquoEf ciency In-vestigation of Scramjet with Detonation and Shockless Supersonic Combus-tionrdquo Utchenye Zapisky TsAGI Vol 23 No 2 1992 pp 30ndash37

37Bruckner A P and Hertzberg A ldquoRam Accelerator Direct LaunchSystem for Space Cargordquo International Astronautical Federation Paper 87-211 1987

38Hertzberg A Bruckner A P and Bogdanoff D W ldquoRam Accel-erator A New Chemical Method for Accelerating Projectiles to UltrahighVelocitiesrdquo AIAA Journal Vol 26 No 2 1988 pp 195ndash203

39Kull A Burnham E Knowlen C Bruckner A P and HertzbergA ldquoExperimental Studies of the Superdetonative Ram Accelerator ModesrdquoAIAA Paper 89-2632 1989

40Patz G Seiler F Smeets G and Srulijes J ldquoStatus of ISLrsquos RA-MAC30 with Fin Guided Projectiles Accelerated in a Smooth Borerdquo Pro-ceedings of the Second International Workshop on Ram Accelerators Univof Washington Seattle WA 1995

41Seiler F Patz G Smeets G and Srulijes J ldquoGasdynamic Limitsof Ignition and Combustion of a Gas Mixture in ISLrsquos RAMAC30 ScramAcceleratorrdquo Proceedings of the Second International Workshop on RamAccelerators Univ of Washington Seattle WA 1995

42Seiler F Patz G Smeets G and Srulijes J ldquoIn uence of ProjectileMaterial and Gas Composition on Superdetonative Combustion in ISLrsquosRAMAC30rdquo AIAA Paper 98-3445 July 1998

43Brackett D C and Bogdanoff D W ldquoComputational Investigation ofOblique Detonation Ramjet-in-Tube Conceptsrdquo Journal of Propulsion andPower Vol 5 No 3 1989 pp 276ndash281

44Yungster S Eberhardt S and Bruckner A P ldquoNumerical Simula-tion of Shock-Induced Combustion Generated by High-Speed Projectiles inDetonable Gas Mixturesrdquo AIAA Paper 89-0673 1989

45Rom J and Kivity Y ldquoAccelerating Projectiles Up to 12 kms Utiliz-ing the Continuous Detonation Propulsion Methodrdquo AIAA Paper 88-29691988

46Humphrey J W ldquoParametric Study of an ODW Scramaccelerator forHypersonic Test Facilitiesrdquo AIAA Paper 90-2470 1990

47Yungster S ldquoNavierndashStokes Simulation of the Supersonic CombustionFlow eld in a Ram Acceleratorrdquo AIAA Paper 91-1916 1991

48Li C Kailasanath K Oran E S Boris J P Paper and LandsbergA M ldquoNumerical Simulations of Transient Flows in Ram AcceleratorsrdquoAIAA Paper 93-1916 1993

49Li C Kailasanath K Oran E S Landsberg A M and Boris J PldquoDynamics of Oblique Detonations in Ram Acceleratorsrdquo Shock Waves Vol5 June 1995 pp 97ndash101

50Yungster S and Radhakrishnan K ldquoComputational Study of FlowEstablishment in a Ram Acceleratorrdquo AIAA Paper 98-2489 July 1995

51Choi J Y Jeung I-S and Yoon Y ldquoNumerical Study of ScramAccelerator Starting Characteristicsrdquo AIAA Journal Vol 36 No 6 1998pp 1029ndash1038

52Bezgin L Ganzhelo A Gouskov O and Kopchenov V ldquoSome Nu-merical Investigation Results of Shock-Induced Combustionrdquo Proceedingsof the AIAA 8th International Space Planes and Hypersonic Systems andTechnologies Conference AIAA Reston VA 1998 pp 72ndash82

53Pratt D W Humphrey J W and Glenn D E ldquoMorphology of Stand-ing Oblique Detonation Wavesrdquo Journal of Propulsion and Power Vol 7No 5 1991

54Fujiwara T Matsuo A and Nomoto H ldquoTwo-Dimensional Detona-tion Supported by a Blunt Body or a Wedgerdquo AIAA Paper 88-0089 Jan1988

55Fan B C Sichel M and Kauffman C W ldquoAnalysis of ObliqueShock-Detonation Wave Interactions in the Supersonic Flow of a Com-bustible Mediumrdquo AIAA Paper 88-0441 Jan 1988

56Dabora E K Desbordes D Guerraud C and Wagner H GldquoOblique Detonations at Hypersonic Velocitiesrdquo Dynamics of Detonationsand Explosions Detonations Vol 133 Progress in Astronautics and Aero-nautics AIAA Washington DC 1990 pp 187ndash201

57Li C Kailasanath K and Oran E S ldquoStability of Oblique Detona-tions in Ram Acceleratorsrdquo AIAA Paper 92-0089 Jan 1992

58Li C Kailasanath K and Oran E S ldquoStructure of Reaction WavesBehind Oblique Shocksrdquo Dynamic Aspects of Detonation Vol 153 Progressin Astronautics and Aeronautics AIAA Washington DC 1993 pp 231ndash

24059Li C Kailasanath K and Oran E S ldquoDetonation Structures Behind

Oblique Shocksrdquo Physics of Fluids Vol 6 April 1994 pp 1600ndash161160Da Silva L F and Deshaies B ldquoNumerical Study of Some Ignition

Regimes of Combustible Supersonic Flows over a Wedgerdquo Proceedings ofSecond International Workshop on Ram Accelerators Paper 15 Univ ofWashington Seattle WA 1995

61Thaker A A and Chelliah H K ldquoNumerical Prediction of ObliqueDetonation Wave Structures Using Detailed and Reduced Reaction Mecha-nismsrdquo Combustion Theory and Modelling Vol 1 No 4 1997 pp 347ndash376

62Desbordes D Guerraud C and Hamada L ldquoSupersonic H2-AirCombustions Behind Oblique Shock Wavesrdquo 14th International Colloquiumon Dynamics of Explosions and Reactive Systems Paper D28 Univ ofCoimbra Portugal Aug 1993

63Li C Kailasanath K and Oran E S ldquoDetonation Structures Gener-ated by Multiple Shocks on Ram-Accelerator Projectilesrdquo Combustion andFlame Vol 108 1997 pp 173ndash186

64Viguier C and Desbordes D ldquoConditions of Onset and Stabilizationof Oblique Detonation Wavesrdquo Advances in Experimentation and Compu-tation of Detonations edited by G D Roy S H Frolov K Kailasanathand N Smirnov ENAS Moscow 1998 pp 40 41

65Bussing Tand Pappas G ldquoIntroduction toPulseDetonation EnginesrdquoAIAA Paper 94-0263 1994

66Helman D Shreeve R P and Eidelman S ldquoDetonation Pulse En-ginerdquo AIAA Paper 86-1683 June 1986

67Cambier J L and Adelman H G ldquoPreliminary Numerical Simu-lations of a Pulsed Detonation Wave Enginerdquo AIAA Paper 88-2960 July1988

68Eidelman S Grossmann W and Lottati I ldquoAir-Breathing PulsedDetonationEngine Concept A Numerical Studyrdquo AIAA Paper 90-2420 July1990

69Eidelman S Grossmann W and Lottati I ldquoComputational Analysisof Pulsed Detonation Engines and Applicationsrdquo AIAA Paper 90-0460 Jan1990

1708 KAILASANATH

70Lynch E D Edelman R and Palaniswamy S ldquoComputational FluidDynamic Analysis of the Pulse Detonation Engine Conceptrdquo AIAA Paper94-0264 Jan 1994

71Lynch E D and Edelman R B ldquoAnalysis of Flow Processes in thePulse Detonation Wave Enginerdquo AIAA Paper 94-3222 June 1994

72Bussing T Hinkey J B and Kaye L ldquoPulse Detonation EnginePreliminary Design Considerationsrdquo AIAA Paper 94-3220 June 1994

73Zitoun R Gamezo V Guerraud C and Desbordes D ldquoExperimen-tal Study on the Propulsive Ef ciency of Pulsed Detonationrdquo 21st Inter-national Symposium on Shock Waves Paper 8292 Great Keppel IslandAustralia July 1997

74Sterling J Ghorbanian K Humphrey J and Sobota T ldquoNumericalInvestigations of Pulse Detonation Wave Enginesrdquo AIAA Paper 95-2479July 1995

75Hinkey J BBussing T R A and Kaye L ldquoShock Tube Experimentsfor the Development of a Hydrogen-Fueled Pulse Detonation Enginerdquo AIAAPaper 95-2578 July 1995

76Schelkin K L Soviet Journal of Technical Physics Vol 10 1940 pp823ndash827

77Bussing T R A ldquoA Rotary Valve Multiple Pulse Detonation Engine(RVMPDE)rdquo AIAA Paper 95-2577 July 1995

78Bratkovich T E and Bussing T R A ldquoA Pulse Detonation EnginePerformance Modelrdquo AIAA Paper 95-3155 July 1995

79Ting J M Bussing T R A and Hinkey J B ldquoExperimental Char-acterization of the Detonation Properties of Hydrocarbon Fuels for theDevelopment of a Pulse Detonation Enginerdquo AIAA Paper 95-3154 July1995

80Stanley SBBurge K and Wilson D ldquoExperimental Investigation ofPulse Detonation Wave Phenomenon as Related to Propulsion ApplicationrdquoAIAA Paper 95-2580 July 1995

81Eidelman S Yang X and Lottati I ldquoPulsed Detonation Engine KeyIssuesrdquo AIAA Paper 95-2754 July 1995

82Aarnio M J Hinkey J B and Bussing T R A ldquoMultiple CycleDetonation Experiments During the Development of a Pulse DetonationEnginerdquo AIAA Paper 96-3263 July 1996

83Pegg R J Couch B D and Hunter L G ldquoPulse Detonation EngineAir Induction System Analysisrdquo AIAA Paper 96-2918 July 1996

84Sterling J Ghorbanian K and Sobota T ldquoEnhanced CombustionPulsejet Engines for Mach 0 to 3 Applicationsrdquo AIAA Paper 96-2687 July1996

85Cambier J L and Tegner J K ldquoStrategies for PDE PerformanceOptimizationrdquo AIAA Paper 97-2743 July 1997

86Stuessy W S and Wilson D R ldquoIn uence of Nozzle Geometry on

the Performance of a Pulse Detonation Enginerdquo AIAA Paper 97-2745 July1997

87Eidelman S and Yang X ldquoAnalysis of the Pulse Detonation EngineEf ciencyrdquo AIAA Paper 98-3877 1998

88Hinkey J B Williams J T Henderson S E and Bussing T R AldquoRotary-Valved Multiple-Cycle Pulse Detonation Engine ExperimentalDemonstrationrdquo AIAA Paper 97-2746 July 1997

89Bussing T R A Bratkovich T E and Hinkey J B ldquoPractical Im-plementation of Pulse Detonation Enginesrdquo AIAA Paper 97-2748 July1997

90Bratkovich T E Aamio M J Williams J T and Bussing T R AldquoAn Introduction to Pulse Detonation Rocket Engines (PDREs)rdquo AIAAPaper 97-2742 July 1997

91Hinkey J B Henderson S E and Bussing T R A ldquoOperation of aFlight-Scale Rotary-Valved Multiple-Combustor Pulse Detonation Engine(RVMPDE)rdquo AIAA Paper 98-3881 July 1998

92Brophy C Netzer D and Forster D ldquoDetonation Studies of JP-10with Oxygen and Air for Pulse Detonation Engine Developmentrdquo AIAAPaper 98-4003 July 1998

93Sekar B Palaniswamy S Peroomian O and Chakravarthy S ldquoANumerical Study of the Pulse Detonation Wave Engine with HydrocarbonFuelsrdquo AIAA Paper 98-3880 July 1998

94Musielak D E ldquoInjection and Mixing of Gas Propellants for PulseDetonation Propulsionrdquo AIAA Paper 98-3878 July 1998

95Sjogreen B and Tegner J ldquoMechanisms Governing DetonationWaves and Their InitiationmdashImplications on the Pulse Detonation EnginerdquoAIAA Paper 99-IS-134 Sept 1999

96Kailasanath K Patnaik G and Li C ldquoComputational Studies ofPulse Detonation Engines A Status Reportrdquo AIAA Paper 99-2634 June1999

97Kailasanath K and Patnaik G ldquoPulsed Detonation EnginesmdashWhat isIts Performancerdquo Proceedings of the 36th JANNAF Combustion MeetingChemical Propulsion Information Agency Laurel MD 1999

98Cambier J L Adelman H G and Menees G P ldquoNumerical Sim-ulations of a Pulsed Detonation Wave Augmentation Devicerdquo AIAA Paper93-1985 June 1993

99Loth E and Loth J L ldquoHigh Ef ciency Detonation Internal Com-bustion Enginesrdquo AIAA Paper 92-3171 July 1992

100Loth E ldquoVortex Formation in a Proposed Detonation Internal Com-bustion Enginerdquo Journal of Propulsion and Power Vol 11 No 3 1995

M SichelAssociate Editor