Design of Surface Combustion Appliances.

OCt.. I 9 1 3 T H E J O l-R-VVA L O F I -VD 1-S T RI A L

unknown, and seizing upon such fragments of t ru th as drifted in lvithin their reach, turned them t o the enrichment of the intellectual and material life of the community. Later they ventured timidly t o launch the frail and often leaky canoe of hypothesis and

A Y D E ,VGI S E E RI S G C H E M I S I' R Y 801

returned with richer treasures. Today, confident and resourceful, as the result of many argosies, anti having learned t o read the stars, organized. equipped, they set sail boldly on a chartered sea in staunch ships with tiering canvas bound for n e x El Dorados.

I ORIGINAL PAPERS DESIGN OF SURFACE COMBUSTION APPLIANCES

By C H A R L E S EDW.4RD L U C K E ~ Received September 10, 1913

l fanufactured gaseous fuel is and always will be more costly per unit of heat carried t h a n natural fuel, and yet may yield cheaper and better service; cheaper i f the process and apparatus used for the combustion of the gas is efficient enough, and better if i t is so designed as t o liberate the heat in a sufficiently more available form. I t is this fact t h a t justifies the general interest now shown in the process tha t has been termed Surface Combustion. which promises both cheaper and better gas sen-ice, the realization of TT-hich depends on the accumulation of much ne\\- da ta for the design of apparatus of commercial form. Some of this in- formation has been worked out , and commercial apparatus of one or tn-o classes designed therefrom. The process of de\-elopment t h a t has resulted in the establishment of surface combustion on an engine-ring basis v-hereby apparatus can now be designed to meet specific conditions is reported briefly in this paper.

A S S U U P T I O X S O F XETY P K O C E S S

The new process assumes t h a t gas t o be burned should lie supplied with no more air t h a n will furnish the required amount of oxygen for the combustion reaction, and tha t the air and gas should be thoroughly mixed previous t o combustion, so t h a t the reaction may take place instantly. once the ignition temperature is reached. Excess air is regarded as not only useless but harmful because i ts heat absorption prevents the at ta inment of the highest temperatures so desirable when the heat of combustion is t o be communicated t o other bodies, and carries away as flue heat. quantities t h a t would otherwise be of use. Premixture is regarded as desirable because thereby each particle of fuel may be brought positii-ely into intimate contact with i ts required oxygen before i t is needed. instead of depending on the accidental dissipation of the products formed on the edge of a flame jet , before the central core of gas can secure air from a surrounding and supporting atmosphere. This premisture. in conjunction with suitable proportioning. prel-ents the escape of unburned fuel in a n y f o r m Thus. premisture of air and gas in combining proportions at once insures protection

inst two important sources of loss in combustion: 1 ) tha t clue t o excess air and ( 2 ) tha t due to incomplete

combustion. .As an incidental accompaniment. another ad\-antage of no less importance follows. and tha t is del-elopment of the heat of combustion in a form more aI-ailable for absorption by the bodies to be heated, and

' Profesqor of 3ltchanical Engineering, Columbia University, New l-ork City.

for the heating of which the gas is burned. Heat will be absorbed most readily from a fire n-hen the temperature of the gases leaving the first is highest and when the fire zone is most radiant. XI1 heat absorbed from the fire b y direct contact of the absorber with the hot gaseous products is absorlieti at :I

rate directly proportional t o the excess of the tcmpcra- ture of the gases over t h a t of the absorber; hcnce. the hotter these gases the more heat will a gi\-en ahsorliing surface take up. other things being equal. Heat is. however. much more rapidly absorbed by liotlies when the source is radiant, because radiant heat readily pierces the insulating dead gas films adhering t o the surface of the absorber and resisting 111- its low thermal conductivity all transmission from passing hot g;is streams. The superior transmitt ing \ -due of radiant heat has been well knoivn as long as physicists hay(: studied the sun's rays. but it has been lacking- in most, though not all. gas burners because of the 1-ery low radiant value of hot gases as compared with solitl bodies a t the same temperature. The premirture of the gas and its supporting air makes it very easy t o secure a large amount of the heat of combustion in :L radiant form. because the combustion, being entirely independent of any atmosphere into which the products may be discharged. can be carried on behind layers of solid granules. in the crei-ices between them. in holes in solid plates, or behind solid plates of any convenient form, all of which. attaining the temperature of the gaseous products of combustion, radiate heat a t a rate immensely superior t o t h a t of the gases themselL-es.

With all these prospectix-e advantages in the direction of superior efficiency of this gas burning apparatus over all other modes of supply, the question naturally arises as to why the principles in\-ol ved haye not been more commonly practiced. and whl- there should be any delay in a t once proceeding t o the design of suitable appliances. The answer is t o be found in the peculiar physical properties possessed by the gas and air over those now commonly in use, which make them difficult t o control in the absence of detailed knowledge of their characteristics. I n fact , n i thout such knowledge, design of apparatus is quite impossible and i t becomes feasible just in proportion as informa- tion of the needed sort is established b y experimental research.

K E Q U I K E 3 I E S T S O P P R O C E S S

The fundamental peculiarity of such mixtures is the property of self-propagation of flame through them. bringing them into the class of things commonly termed esplosi\-e. Xs all the mixture is in a condition suitable for combustion once the ignition temperatiirc

8 0 2 T H E J O U R N A L OF I N D U S T R I A L

is reached, it naturally follows tha t as t he combustion temperature is much in excess of t ha t of ignition, t he

a t a point of ignition will promptly heat neighboring layers of mixture far above their ignition temperature so tha t t h e flame will, of itself, proceed through the whole mixture mass if it be isolated in a chamber. The only way such a self-propagation can be stopped is by preventing the heating of a fresh layer by the combustion of i ts neighbor, such prevention taking the form of a heat absorption by some solid screen a t a rate great enough t o permit the screen to take up the heat of combustion of all mixture in contact with one side without itself becoming hot enough t o warm the mixture on i ts other side t o the ignition point. It is easy t o make such flame interrupting screens tha t will stop the propagation of flame through a n explosive mixture once or twice, bu t difficult or impracticable t o make them so as to be able to continue to do so indefinitely. With the exception of the internal combustion gas engine, gas burning apparatus requires a continuous burning of gas as supplied in a definite place termed the burner or furnace. When explosive gaseous mixtures are supplied t o such continuous fires, t he property of self-propagation operates t o defeat localization of the combustion unless specific means are provided in the design of the structure t o make the treatment of t he mixture conform t o its physical properties, in which case the combustion can be more definitely localized than otherwise.

For any given mixture there is a definite normal rate of propagation which may be equal to, less or greater than the velocity of flow of the mixture into the fire. and i t is clear t ha t localization of combustion will depend primarily on the relation of these two rates. t ha t of flame propagation and tha t of flow. If they are equal, t ha t is, if the flame can burn back toward the source of mixture supply, just as fast as, but no faster than the mixture reaches the flame, then will the flame remain fixed or be definitely localized. I n fact, the above is the primary condition for localization of the combustion of explosive mixtures. because otherwise the flame would travel back to the source of supply or be blown out by the physical pushing away of t he faster flowing fresh stream of mixture. Of course, localization may also be secured if the mixture be fed through a flame-interrupting partition a t a flow rate inferior t o tha t of propagation, in which case the localization takes place on the face of the partition, which must, of course, have a large capacity for heat absorption indefinitely, i. e., i t must be able t o dissipate or transmit t o some other body the amount of heat i t is receiving-otherwise i t would rise in temperature to the point where i t , by igniting the mixture on the supply side, would cease to be an interrupter.

Mixtures of the sort t ha t seem best from the stand- point of efficiency of apparatus, having this explosive property, require a special treatment t o permit of their use in commercial apparatus, which special treatment has for i ts first object the definite localization of the flame. The means employed must, moreover, be so positive as t o be unaffected by long-continued operation;

\ burning of however minute a quantity of mixture

A-VD E;VGINEERIAVG C H E M I S T R Y v01. j. SO. I O

t h a t is, t he localization must be permanent. There are, however, still other conditions to be met, such as control of rate of combustion per sq. f t . of fire or per CU. ft. of furnace, a cook stove requiring a low and a crucible furnace a high rate, and as such mixtures have what might be called a natural rate corresponding to the normal rate of propagation, it is clear t ha t specific means must be available for burning slower or faster per sq. f t . of gas stream cross section. Furthermore, any one burner or furnace designed for some definite or normal rate of combustion or gas consumption, would be valueless commercially unless it would operate quite as positively a t a wide range of variation, both above and below tha t normal rate. I n other words, localization of the flame must not only be positive, and any desired rate of combustion be attainable by design, bu t in addition-hand or automatic interference with the designed rate of consumption must not in any way interfere with the localization.

I n some cases i t is desirable tha t t he burner be capable of reaching i ts normal or steady working s ta te in a short time, e. g., in a domestic cook stove burner, while on the other hand a slow heating corresponding to a large heat storage in the furnace is necessary when articles are constantly being thrust in t o be heated and then drawn out, without too much change of temperature in the furnace: hence, control of heat capacity of the burner must be available. Finally, there must be provided mechanical means for making the desired mixtures and for maintaining the desired proportions sufficiently, means so simple and automatic t o require no more skill or attention on the par t of the operator t han appliances nomy in use.

Accordingly, this review of the work of development of the first commercial surface combustion appliances is divided into the following divisions:

I. Localization of Combustion Zone: i n i t i a l l y on starting cold apparatus and pevmanelz t ly on attainment of the steady state of t he fire.

11. Ra te of Combustion per sq. f t . of Bed or Hearth for High Rates and Low. 111. Control and Adjustment of Radiating Surface. IV. Auxiliary Apparatus. V. Efficiency of Surface Combustion Appliances. The experimental work here reported has all been

done by or under the direction of t he writer, partly in his laboratory a t Columbia University, but largely in the laboratory of t he Gas and Oil Combustion Co. in the Chemists’ Building. Acknowledgment is freely made of the contributions of t he laboratory staff, headed by Mr. Frank Creelman and including Messrs. E. J. Allen, H. L. Ocumpaugh, F. A. Wegener, W. B. Eddison and Prof. E. J. Hall of Columbia University.

Only the results of S . Y. City illuminating gas are here reported, though work has been done with other fuels and satisfactory results obtained tha t permit the statement t h a t all fuels in the gaseous or liquid form are equally available for the process.

I. L O C A L I Z A T I O N O F C O M B U S T I O N Z O S E

Reasoning from the fact t ha t the flame cap or surface over which combustion is proceeding will be steady in

OCt., I913 T H E JOCRAVAL O F I -VDUSTRIAL

location only when the velocity of t he mixture ap- proaching i t is equal t o the ra te of flame propagation for such a mixture, the first step in t h e design of prac- ticable apparatus is the formation of gas passages of suitably changing area, since the velocity of gas flow for small changes of pressure will vary inversely as t he cross section of t he flow pa th . The mixture should approach the fire zone through a small tube or orifice a t a ra te greater t han the normal velocity of flame, bu t immediately beyond there should be a n enlargement of t he cross section of the passage t o permit of velocity reduction which, if carried far enough, would finally cause flow velocity to drop t o equality with flame velocity, which was a t t he point of entrance in excess. Mechanical construction tha t will permit of this can be designed in almost endless variety, many of them equally good, bu t for the best results certain precautions must be observed: notably ( I ) the tendency for a gas jet escaping from a n orifice under pressure t o resist sidewise expansion so t h a t even if the passage offers an increasing area for flow, the gas stream may not fill i t unless forced to do so by baffles or their equivalent; a n d ( 2 ) t he tendency for free jets flowing through a gaseous atmosphere of, for example, products of combustion, t o entrain or diffuse with the atmosphere at the edges, becoming so dilute as t o not burn properly, if a t all.

GAS P A S S A G E S

Perhaps the simplest apparatus would be a very narrow angle conical passage to which mixture is fed a t the smaller end ; the stream expanding sidewise t o always fill the cone would reduce in velocity and permit t he location of a combustion surface quite definitely. I n time such a conical tube would get ho t ; in fact. a t t he section where the flame surface located, it would become incandescent almost immediately if of refractory material, or melt off if of metal, and, of course. t he narrower section along which the fresh mixture approaches would heat by conduction and warm the feeding gas. This mixture heating would tend t o increase flow velocity and push away the flame surface if another counter influence did not prevent i t ; i. e . , the increase of flame propagation rate, which is due partly to rise of mixture temperature and partly to the accelerating effect of the hot solid walls. This counter influence accelerating combustion and drawing the flame surface back against the supply current is really stronger than the pushing away effect of velocity increase due to expansion of the stream by heating. Such cones are not in most cases good forms for t he purpose because of t he limit imposed on adjustment of rate of supply which if varied very much would cause a blow off or back flash of t he flame, unless they were extremely long. When, however, a steady- ra te apparatus is needed and flow is fixed, there may be just enough sidewise expansion of t he jet issuing from a mere hole in a plate t o locate the flame, bu t t he practical uses of such apparatus are very small.

To meet t he needs of every-day service there must be a more positive localization of the combustion surface, insensible to variation of mixture pressure or flow rate over a wide, though definite service range;

A ND E LVGI S E E R I N G C H E i?fIS T R Y 803

this requires specific provision for stream baffling so t h a t t he gas stream will increase in cross section rapidly beyond the supply point of minimum area and safe maximum flow velocity. Perhaps the simplest of these flame spreaders is a plate over the hole through which the mixture escapes as

I

ROU F SQU A

in Fig. I . By placing

2

FIGS. 1 TO &MEANS FOR TELOCITY REDVCTIOX TO CONTROL LOCALIZA- T I O S OF C O M B V S T I O S SCRf .%CE

t he baffle away from the plate carrying the orifice, a distance equal t o one-quarter of the diameter of t h e orifice, the stream direction will be turned through a right angle without changing the flow velocity until the stream moves out under the plate beyond the hole, when, of course, t he velocity will be reduced inversely as the diameter of the circle the gas occupies a t any one time. With a hole 1/16 of an inch in diameter a plate spaced of an inch away could reduce flow velocity when cold to l / g of its entrance value if the plate were

This would permit an eightfold change in the rate of flow when cold, keep the flame surface always under the plate and give to the flame surface the form of a short cylinder. Such a construction offers some difficulties as in operation both plate and hearth would get very hot and unequally so; starting at the center the plate would remain dark, becoming incandescent from the flame zone out t o the edge, which would crack most plates refractory enough to resist fusion even if a practical way of making such plates and holding them in position were available.

An easier construction is shown in Fig. 2 , in which a solid (alundum cement) cone with three small ribs is placed in a counterbore of the feed hole. This is self-locating when in the position shown but not otherwise. Still another form tha t holds the baffle definitely against side movement is shown in Fig. 3, in which the baffle is shaped somewhat like a mushroom with a square-section stem fitting loosely in a n enlargement of t he feed hole so t h a t the flow area does not change materially until the escaping mixture strikes the underside of the head and is deflected outward. The dimensions given are those used on some tha t gave good service with mixture pressures up t o three inches of water though a little troublesome t o make and ap t t o break if not gently handled. The most permanent form of this class of molded baffles is the refractory ball in a conical hole as shown in Fig. Z+ and i t is inter- esting to note tha t the ball, remaining suspended in

inch in diameter.

,

T H E J O C R N d L O F 1,VDCSTRIAL A J D E S G I S E E R I S G CHEMISTRY-T'ol. j . S o . IO

the jet and keeping, b y reason of i ts weight, a constant velocity of the jet around it . insures the location of the flame above i t ? keeping i t cooler than otherwise, so t h a t fire clay balls have remained uninjured when the same clay a short distance above readily fuses. The physical action is here slightly different t h a n in the previous cases, the escaping mixture forming an expanding and velocity reducing jet of ring form between t h e ball and i ts seat.

While. in the construction shown as typical, localization of the combustion surface is definite and mixtures may be burned in combining proportions, all fail t o produce the uniform incandescence tha t is easily obtainable otherwise with such mixtures, though. of course, a series of such baffles spaced uniformly, approach this result; instead of a uniform'glow they have a spotty appearance from black t o truly incandescent. I t is found, however, t h a t any sort of porous screen placed in the pa th of the escaping gases will get uniformly hot. Such screens have been used of three sorts: ( I ) .% perforated plate of metal: made of one of the non-oxidizable alloys such as are used in electric heaters. ( 2 ) ,4 layer of granulated refractory material. (3) ,4 diaphragm of bonded granular material, the most satisfactory of which were made of S o r t o n alundum. Of course. the less mass given t o the screen the more rapidly i t reaches incandescence initially and in this respect the metal screen is best; one of these placed '8 inch over a hearth 4l I ? inches diameter supplied with j j holes each 16 inch diameter and j / ' 4 inch ball baffles heated in six seconds, while diaphragms thi,ck enough t o be mechanically strong may take five minutes, though, of course, the latter are more permanent. With the same arrangements except t h a t the mushroom baffles were used instead, the following consumptions of gas were measured a t the corresponding pressures and gave a steady dull red a t the lowest ra te and correspondingly brighter heats at the higher rates: j . 2 5 , 9.1, 10.9, 1 2 . 2 . and 1 4 . 2 cu. ft. of gas per hour a t 1.00, I . j , 2 . 0 . 2 . j and 3.0 inches of water pressure. A similar set of mushroom baffles placed in a vertical hearth and having a screen of crushed fused silica inch deep of size between ":64 inch and 1 3 / 6 4 inch, held in place by a second screen of mire mesh, heated t o incandescence in 3l/2 minutes a t a supply pressure of 2 inches of water.

A11 the previously described baffling means intended t o assist the spreading of the mixture stream beyond the point of high velocity supply, and many others of the same class, are used t o counteract the natural tendency of gas jets t o remain intact and resist sidewise expansion. It is possible. howel-er. t o make use of this natural tendency of jets and incidentally secure a very high degree of incandescence, higher in fact t h a n is possible by any other known means, using- air and the same gas. Moving the nozzle away from the baffle and making the baffle porous (using for canve- nience in experimental xyork a pile of refractory granules), the top or impact surface may be shaped t o conform t o the natural mushroom of the jet as it strikes. For any degree of porosity and jet velocity or supply pressure there is a best distance and a corresponding

impact face shape, tha t permits all mixture striking whether a t the center or edges t o have just the right velocity, t h a t of propagation, in contact with the material. so t h a t combustion takes place in contact with the outer or radiating surface which then has the maximum possible temperature. I t is essential tha t the impact face be porous so t h a t the products may

F I G . s--hIEASS FOR T,OCALIZING THF: C O M B U S T I O N Z O S E -42 THE EXTERNAL RADIATIXG SYRFACE

readily-escape through instead of forming dead gas films which separate the combustion surface from the solid faces and insulate the latter from the heat of the former. In this manner a surprising degree of radiance is obtained due t o the high rate of combustion of a mixture containing no excess air except possibly a t the outer edges by entrainment. At 3 . 2 inches mixture pressure, 40 cu. f t . of gas per hour were burned over about four square inches of surface, which corresponds t o over 1400 cu. f t . per sq. f t . per hour and more t h a n 800,ooo B. T. U.'s per hour per sq. f t . Higher rates of combustion than this are obtainable but not all developed on the radiating surface itself.

P R O P E R T I E S O F G R A X U L A R B E D S

I n view of the importance of developing as large an amount of heat as possible in the radiant form, in addition t o the localization of the combustion zone, and considering t h a t loose granules or a porous diaphragm of refractory may serve the double purpose of the radiating screen and of the actual baffling or mixture stream spreader, everything else may be abandoned in favor of these refractories except perhaps for certain special purposes for which special forms may he best. Nearly all the work done has been concentrated on an effort t o learn how these granular beds and bonded diaphragms should be handled t o yield commercial apparatus, and b y reason of the simplicity of the loose bed i t has been consistently favored for most uses over the bonded. X uniform grained bed of loose granular material

\\-ill act as a substantially perfect spreader of the mixture issuing from an orifice covered by the bed but the quantitative effect of variations in size of' grain, imperfections in shape or non-uniformity of size, together with shape of bed and its relation to shape, size and position of the feed orifice, are all t o be de-

OCt.. I913 T H E JOL7RLVdL OF I J Y D Z - S T R I A L ilLVD E S G I L V E E R I S C C H E M I S T R Y S O 5

termined b y experiment only. N o s t of t he detailed I n t ime, however, t he lower side of t he top layer gets d a t a of this sort will be omitted as not of general hot and the combustion surface recedes, doing so interest bu t t he following series of experimental de- more rapidly once it has passed under the top layer terminations illustrate t he most impor tan t of t he because then radiation cooling is lessened, leaving general principles t h a t must be observed for successful more heat for t he grain to take up. I t follows also tha t operation: One of t h e simplest imaginable arrange- as with the finer grain less total mixture will escape ments is shown in Fig. 6 , where a single nozzle j/16 from such a round nozzle as i t tends to fill t he orifice,

FIG. 6-APP.4RATUS F O R DEMONSTRATISC THE SPHERICAL FOR= OF THE

COMBVSTION SURFACE WHEN MIXTURE IS SUPPLIED FROM A NOZZLE SURROUNDED BY GRAKULAR REFRACTORY

inch in diameter projects 1 ~ ' 4 inches into a bed of broken fire brick which material is very convenient for studying the ult imate position and shape of t he combustion zone as i t will fuse and bond a t t ha t place. The bed was held in a square brick box 63.'14 inches on a side and j inches deep and t h e nozzle was fed with mixture a t 3 . 2 inches pressure. Start ing with coarse grains l,,'? inch t o 3 / ~ inch. t h e mixture on ignition flashed instantly down through the bed, locating bel.ow, quite near t he nozzle as was shown later by the more or less sp,herical lump t h a t fused there. The same action took place with smaller grain as it was reduced in size successively from "8 inch t o inch, 3/16 inch t o ' /4 inch. l j 4 inch t o 3/64 inch, inch t o 6,/3? inch and 5 / 3 2 inch t o 3 /32 inch, bu t a further reduction in size resulted in a different action. On using grains f rom 3 , i g ? inch to '/I6 inch in size, t he mixture on ignition

so t h e total amount of heat developed will be less t han with coarse grain. To show tha t t he stoppage of t h e nozzle by the grain reducing the total gas flow and heat generation with fine grain is not responsible for t he difference in flame action a t the s ta r t as compared with coarse grain bu t t h a t t he difference is due to the difference in the mode of gas flow through the bed itself with corresponding more intimate contact be-

l l

I FIG i--dPPARATUS FOR DEMOKSTRATING c N I F O R M I T Y OF POSITION OF

COMBUSTION SURFACE WITH DIFFERENT SIZES OF GRAIN

tween mixture and grain t h e smaller t he latter, the same series of results were obtained in a pyramidal chamber, shown in Fig. j , in which the nozzle is a pair of hacksaw cuts in a I-inch pipe cap, t he whole being shown in the photo. Fig. 8. With such slot outlets for the mixture the flow is substantially the same for all t he different sizes of grain given, yet not only did the top heat first with t h e 3 / 3 2 inch t o inch

- burned on :op instead of flashing down through the voids T O i:s normal surr'zce about t he nozzlr. In a few Ini2ures :he T O P layer 5ecame hot and then seemed t o cool: b u t on scraping a w a y :he top it coulc! Le seen t h n t *.he xmbus r ion jurr'::cc hzd receded: nr'xer r-.inning a sufrji-i-nt time :his progressive heating from the out - sir!c (!on-n roward the nozzle r e s u l r c d in ;i fin:,l locarion of t h e combustion zone ~ ~ o . . I T where i t wns in the co;trsr'r irx:eri31. in w h i c ~ i i t flashed bzck i ; i j f c i i r t / x

t o i!s ncrinn: position. Or' course. in ci..iicr c~::se. tke whole ; ccl gcrs hot in t ime. bu,: wirh c o i r x gr:,ins i t

i:om :!:e r?orr?xl comLr.srion zone ne:tr t h i nozzle out and upward, while with small enough grain the outside top heats first. t h e flame drawing down slowly, a difference in action which requires a n explanation.

K h c n the grain is fine enough it acts like a flame interrupting screen. each little crevice betlq-een neighboring grains discharging so little mixture tha t its heat of conilmstion can for a n appreciable time be absorbed by t h e solid grain in contact v i thou t raising the lower side to the ignition point, radiation from the outside surface helping to keep down the temperature.

FIG. 8-.APPARATUS FOR DEMOSSTRATIKG c S l F O R X l T Y OF POSITIOl i OF

COMBUSTIOS SURFACE WITH DIFFERENT SIZES OF G R A ~ N

grain and then burn back but also with t h e larger inch to 3 / 3 2 inch with which i t required ten minutes t o draw the combustion zone down through the bed t o i ts normal place and fuse a lump of granules over t he cross slot. In establishing the action just described the use of

fire hrick for t he grain was most satisfactory as each

8 0 6 T H E JOIIRX.4L O F I S D I - S T R I A L

test gave a fixed record (in the form of a fused lump) of what had happened, bu t once it was clear t h a t t he size of grain made only a start ing or temporary difference in action, then longer runs became desirable t o bring out t he possible heating effect on the nozzles; for these refractory grain was used, including magnesite, chrome ore, both natural lumps and broken brick, fused silica, lime and alundum. With these materials t he long runs showed t h a t some forms of feed orifices became heated so as t o ignite the mixture in the feed pipe while with others this action did not take place, permitting operation for a n indefinite time. I t was not clear a t t he time just what caused the trouble since i t frequently happened t h a t t he nozzle t h a t worked best was really much hotter t han another t ha t flashed back almost as soon as the steady s ta te was established. At any time an increase in the pressure at which the mixture was supplied would have so increased the volume of flow as t o drive t h e combustion surface to a safe distance from the nozzle bu t i t was deemed unwise t o leave the matter here because, f i r s t , all domestic apparatus must operate on low pressure, lower than t h a t of the gas in the local mains, and second, all industrial apparatus for which gas pressure boosting would not be objectionable would, a t some time, be operated a t par t capacity or turned down, which is equivalent t o low pressure operation.

PERMANENCE OF FLAME SURFACE

The next question investigated, therefore, was this permanence of the localization of t he flame surface or combustion zone, and the discovery of means for securing an indefinite t ime of operation without disturbance from long-continued heating even over several days and with mixture pressures so low as t o heat t he hearth end of the nozzle above the ignition point. It is clear a t t he s tar t t h a t iron, or in fact a n y metal, must be kept away from the feed orifice proper, which in effect means t h a t in order t h a t appliances may operate on low pressures i t is necessary t h a t the mixture enter t he fire chamber through a refractory orifice, or when the burner is lined with refractory, as i t should be, t he mixture must enter through a hole in the lining or what might be properly called t h e hearth. Using different mixture entrances a long series of experiments were made t h a t were for a t ime extremely puzzling bu t which finally could be summed up in a very few words. All feed holes will become incandescent a t t he fire end with low gas pressures; heat $11 flow back through t h e bounding material of t he hole and these hot walls will heat the mixture passing through. N o harm will result, however, even from a bright red heat on the walls of t he hole, provided the latter is of uniform bore over the heated par t and has no enlargement of cross section with i ts corresponding low mixture flow velocity except a t points t h a t never get hotter than the point of ignition. T o pu t i t otherwise, the mixture passageway may be a t any temperature without harm if of uniform bore and if when cold the flow velocity is great enough to prevent back flash. Under these conditions localization of the combustion will be permanent . Naturally anything t h a t will reduce the tem-

A-VD E S G I S E E R I - V G C H E X I S T R Y T'ol. j. SO. I O

perature of the supply tube, like radiation to the air, will also shorten the length of the necessary uniform bore portion as will also the use of more non-conducting material. Finally i t was found t h a t the length of uniform bore feed pipe t h a t would be cool enough a t its supply end t o be joined t o a larger mixture header or box, was greater for large than for small holes or t h a t the allowable length was a definite number of diameters. For example, in an apparatus such as is illustrated in Fig. 9, with a single refractory lined tube - .

feeding a cylindrical combustion chamber, all of a lundum cement encased in metal, a hole of l / 4 inch diameter and 6 inches long could be operated indefinitely a t I inch water pressure even after the chamber had been made very hot by previous operation a t much higher pressures, whereas a 7/16 inch diameter hole of the same length flashed back a t 1 . j inches water pressure, and continued t o do so after periods ranging about three hours even after i ts length had been increased to I 5 inches. Small diameter orifices of say 1/16

inch diameter have, on the other hand, been operated indefinitely on pressures as low as 0 . 2 inch of water when about I ~ / Z inches long. Back flash after long periods of operation by heating up of the supply system may, therefore, be prevented by proper design; i e . , by selection of suitable proportions in accordance with the announced principles and the established

FIG. APPARATUS FOR DETBR- MINING THE RATIO OF LENGTH TO

CONTROL "BACK HBATING" DIAMETSR OF FEED HOLE TO

data , localization of the combustion zone will be correspondingly insured against disturbance. 11. RATE OF COMBUSTION PER SQ. FT. OF BED OR HEARTH

Passing now from the discussion of localization of t he combustion zone or flame surface and the means for keeping it permanently localized, t he next question of importance is t h a t of control of rate of heat generation per sq. f t . of fire bed, which is a problem of distribution of t he burning mixture over surfaces as large or small as may be demanded and of maintaining high or low degrees of incandescence over these surfaces. This now seems a very simple matter bu t for a considerable t ime it seemed even more elusive than the prevention of back flash through nozzles, which was annoying enough. There are really two problems here and though related, t he relation is not as clear as it may seem. First , t he problem of the high rate-At how high a ra te may gas be burned and how may the

OCt., I913 T H E J O 1-RS.4 L O F I-VD 1-S T RI.4 L A S D EAVGISE E R I S G C H E M I S T R Y 8 0 7

structure be formed so as t o permit of steady supply and combustion a t this rate over any given surface? S e c o n d , the problem of the low rate-Can such mixtures be burned so slowly as t o keep a surface just hot enough for fusing for example; may such a burner be turned down to a so-called simmering point without giving trouble and if so, what ideas must be incorporated in the construction?

HIGH R A T E OF C O J I B V S T I O X

The problem of the high rate is the simpler of the two, as this may properly be termed the natural rate, assuming, of course, tha t suitable refractories are obtainable for the corresponding high temperatures and t h a t the principles of permanent localization of the combustion zone are understood and incorporated in simple constructions of apparatus. When the whole cross section of bed is taken up by the burning gas stream or when the combustion zone occupies the whole of the bed cross section, then mixture is being supplied as €ast as i t can be burned within the space available and any increase in rate of supply mill tend t o push the combustion zone away or result in the phe- nomenon of blow off. This is the condition a t the limiting high rate of combustion and is the natural rate, because the advancing stream of mixture fills the whole available cross section of fire zone and combustion surface coincides with bed cross section in area. T o establish this high rate i t is necessary t h a t a t first when everything is cold, the mixture be supplied a t a lower rate and then after the bed is heated, its accelerating effect may be relied upon to hold localized a very much increased rate of mixture supply up t o a certain maximum, which is the natural or high rate of combustion. This natural high rate is intimately associated with the pressure of the mixture a t the point of supply, or more properly, the drop in pressure through the feed holes or orifices, for holes of any given size. For example, in such a simple apparatus as t h a t of Fig. 9, where a single feed hole supplies a combustion chamber of larger cross section, carrying a loose granular bed, it is evident t h a t a t some low pressure the combustion surface will just cap the feed hole and the rate of combustion will be natural for a bed area equal, not. to the actual bed of z 3 t 4 inches diameter, but on the contrary, equal t o tha t of the feed hole, a fraction of an inch in diameter. Increase of mixture supply pressure will push the combustion surface away from the hole until finally it will extend across t h e whole bed and have a form approximating a section of a spherical surface lying within a cylinder. the avis of which passes through the center of the sphere, which latter will be located a t the center of the feed hole. I t is this latter condition t h a t gives the high rate and t o secure i t the supply pressure must be big enough t o send through a small feed hole, say 1/4 inch in diameter, enough mixture t o spread out and fill the voids in the bed z ~ / / ~ inches in diameter, which will exceed half the free cross section, while a t the same time the flow velocity through the enlarged cross section must be as great as the rate of flame travel backward. If the ratio of bed area (with a giren percentage of voids) t o feed hole area is large, then high mixture

pressures must be used t o secure the natural high rate; so if only small mixture pressures are available then the free flow area of the bed may not exceed the area of the feed hole by very much. I n any case large beds may be heated by the high rate of combustion, only by multiplication of orifices spaced a t distances equal t o the diameter of the combustion surface natural t o the supply pressure in use. Before such spacing of feed holes for high natural rates of combustion can be found, i t is first necessary t o investigate the action for a single hole in a given bed, t o find the relation between feed pressures and the ratio of active bed area t o feed hole area tha t will cause the combustion surface t o fill the whole bed between the enclosing walls. As an illustration of the method of procedure, i t being impossible here t o record much of this sort of data , the results of a single series on the apparatus shown in Fig. 9 will serve. In this series the pressures were never extended above 1 2 inches of water a t the point of mixture supply and began with a single hole " 1 8 inch in diameter, discharging into a cylinder 2 3 / 4 inches in diameter, area 6 sq. inches, b u t as increase of pressure t o the maximum noted failed t o supply as much mixture as could be consumed in the available area, there was added a second, a third and finally a fourth hole, even then not reaching the limit.

F I G IO-RATE OF F I , O % OF ?IIIXTURE T 1 i K O U ~ ; I I DIFFEREST N U M B E R S O F

5/16" FEED HOLES AT VARIOUS PRESSURES

With the four holes and a t a pressure of 1 6 ~ / 2 inches of water, there was attained a rate of combustion consumption of 2460 cu. ft. of gas per sq. f t . of bed per hour, which on a calorific value of 600 B. T. U. per cu. ft. is equivalent t o 1,j84,000 B. T. U. per hour per sq. f t . There was no indication t h a t the limit had been reached, bu t as the rate was high enough for practical purposes and so high as t o give trouble with available refractory materials, this series was not carried beyond these limits, though i t could easily

808 T H E J O r R . V d L O F I X D L - S T R I A L d,VD E-VGISEERING C H E M I S T R Y Yol. j . S o . IO

have been done with high pressures or with a larger feed area than from ' / I 6 inch hole i= 0.3 sq. in.). The conclusion here is, t ha t rates higher than have been known are easily obtainable and rates higher than any refractory in common use can resist; hence, if any service arises tha t demands such rates, t he need can be met. It might be noted here tha t of several refractories tried, the most satisfactory found was white alundum, with linings of alundum cement. preferably mixed with alundum grain of several sizes, properly proportioned dry to fill the voids, then mixed wet, rapidly dried and finally baked a t a bright red heat. The size of grain in the loose bed and the depth of bed each exert an important influence within limits bu t in general for these high rates the larger t he grain the better, though there is little, if any, advantage beyond one inch diameter, the depth in any case being little more than sufficient t o prevent the gases lifting the grain off the hearth. The resistance in the bed is always appreciably less for round grain than for flat and in this respect the form of fracture of white alundum is most satisfactory. In the above mentioned series white alundum from * '2 inch to '4 inch diameter was used and the depth averaged z inches. The results of this series are given in curve form, Fig. I O .

L O W R A T E O F C O M B U S T I O N

Probably the most difficult of these experimental problems was tha t of finding a suitable structure for ' the permanent localization of the combustion surface for very low supply pressures, while at the same time securing any desirably low rate of combustion over the bed with a uniform incandescence a t the surface such as would be suitable for cooking, for example. It is clear t ha t the lower the pressures the closer t o the feed hole will the combustion surface locate, the smaller t he area of bed heated, and the stronger the tendency to back flash. This fact led to early attempts to secure low rate heat distribution by other means than t h a t of feed hole supply with a series of spots, each having natural high rates of combustion, located practically a t the mouth of the holes. Mixture supplied t o a passageway filled with fine granules or a porous brick, will flow much the same as through a pipe as the porosity supplies a multiplicity of small paths more or less mutually communicating, and a given quantity flowing would have a mean velocity through such a porous layer the same as through a pipe of cross section equal t o the voids except for differences in friction. Mixture fed into such a fine bed in any convenient way, ranging from a single hole under a thick bed to a great number of holes under the whole bed obtainable by a metallic screen support for loose granules or the under surface itself of a bonded diaphragm, would discharge the mixture uniformly from the top of the bed, however the supply was arranged, if the bed were thick enough. This seemed t o be a t first the most promising idea to follow for low rate heat distribution especially as changes of area of flow pa th so necessary for localization can be obtained by two means quite conveniently: j i r s f . by using two layers with the upper coarser and. therefore, with more free voids than the lower fine bed; second, by changing

the diameter of the containing walls t ha t hold the porous bed.

At first only loose granules were used in apparatus such as shown in Fig. 11, where a perforated metal plate of 306 holes per sq. in., diameter 0.033 inch, and 2 7 . 8 per cent free opening is supplied by a inch

7-

FIG. 11-1IEANS OF CONTROLLING SURFACE DISTRIBUTION OF HEAT BY h f I X T U R E SUPPLY TO FINE GRANULAR BED O N T H E SCREEN

pipe, discharging into a conical space under the screen. On the top of the screen rests a one-inch bed of granules

inch to 6 / 6 4 inch and over i t another one-inch layer of larger grain, 2 4 / 6 4 inch to 20/64 inch diameter. On lighting the mixture supplied a t z inches water pressure the combustion zone instantly locates between the layers a t the point of change of free flow space, due t o the difference in voids, bu t as the heat penetrates t h e lower or fine bed, t he combustion zone works back t o the screen, which happens with these conditions in about ten minutes. All sorts of changes in the relative sizes of the material and thickness of the two beds or number of layers yield no difference in the result except as t o the time i t takes the combustion zone ultimately to work back t o the screen, burning i t out, causing a flash back, or both. This is $early due to insufficient velocity of flow a t t he supply end of the bed and suggests a contraction of t he walls toward the lower end, as in Fig. 1 2 , where a slot supply is substituted for the screen as i t d6es not burn out so readily if by accident the flame should work down tha t far. As before, however, the combustion zone worked down close t o the nozzle but not so quickly as before, sug- gesting a further narrowing a t the bottom to form a neck as in the forms successively tried, and shown in Figs. 13. 14 and Ij. As the neck is thus narrowed the combustion surface no longer approaches the nozzle so closely. stopping, in Fig. 13, about I inch above after 3j'4 hour a t 3l/2 inches pressure, in Fig. 14 about 1l/2 inches above, while in Fig. ~j i t did not enter the lower cylindrical portion a t all.

By adjusting the space between the walls, the flow velocity through the lower or fine bed can, therefore, be kept high enough to control the location of the combustion zone, regardless of the manner of feeding

OCt.. I913 T H E JO17RS. - IL O F I S D I - 5 TRI . IL -1ND E X G I S E E R I S G C H E M I S T R Y 809

the mixture t o its narrow bottom part , but the problem is by no means solved. A l l these latter forms are bad

+--- 42 - u-q;-

I 2 13

F I G S . 12 T O lj-!%PPAR.4TUS FOR CONTROLLING POSITION O F C O M B U S -

TIOX SURFACE FOR A GIVEK HEAT DISTRIBUTIOK OYER THE RADIATING SURFACE

in-another respect because with the thickness of bed above the combustion zone the top is hot in the center and cool a t the edges. T o get an even top heat, a greater thickness of bed above the combustion zone would be required but this is already too thick t o give a quick initial heating, so another change in construction is suggested, incorporating a different relative position of the narrow lower neck of fine material in the top surface maintaining about the same relative areas as in the last case, which were found sufficient t o hold the combustion zone above the cylindrical portion. Such an alteration was made by enlarging the cylindrical portion and inserting a solid core 3l/2 inches in diameter so the mixture rose in a thin ring

inch thick around the core filled with fine material and then spread out a t the top, heating a bed of j inches diameter. This gave a better top heat distribution but took too long t o reach a steady state-almost an hour. After many trials of various proportions it became clear t h a t if uniform top temperatures are t o be quickly obtained with low rates of gas consumption over an extended surface t h a t single feeds through single or double porous beds will not suffice and some other plan must be employed.

I

USE O F M U L T I P L E P E E D O R I F I C E S

The failure of the single point supply to distribute mixture so as not t o flash back and yet quickly and uniformly heat the top of a radiating bed. with a low rate of combustion, leads t o the conclusion tha t mul- tiple feed orifices must be used, so spaced as t o give the heat distribution desired a t the available pressure. a surface bed or radiating screen of sufficient thickness serving to localize the combustion and t o equalize the otheryise spotty appearance. I t might appear tha t this could be done by drilling holes in a plate and dividing the combustion chamber from the mixture supply chamber, but from what has been said about the conditions for permanent localization of the combustion zone. i t is clear tha t such an orifice plate will sooner or later heat through, becoming hot enough on the supply side t o ignite the mixture there before i ts entrance into the feed holes. Any sort of feed

hole plate is out of the question unless provided with means for removing heat received on the combustion chamber face, fast enough t o balance heat conduction through it toward the source of supply, and effective enough to prevent tha t side from becoming a point of ignition.

Since i t has been shown tha t a single long feed hole supplying a combustion zone can be proportioned so as t o permanently resist backvl-ard heat flow well enough t o insure indefinite periods of operation. it naturally follows tha t several feed holes might be grouped so as t o be fed from a common point. discharging mixture a t the other end over as large an area as needed for any desired surface rate of combustion, and so proportioned as t o resist heat flow as effectively as with one. if the necessity for the latter is kept clearly in mind. This conception is the basis of the construction shown in Fig. 16 which has 30 holes molded in alundum cement b y wires held in hole plates a t the end and radiating from a inch bottom chamber in the iron container casting, t o a 31;'4-inch diameter circular top bed. T o insure the supply chamber end of these feed holes against a rise of temperature t o the ignition point, it is necessary only tha t the distance between

FIG 1 6 - F l ~ s T COOK STOVE T O P BCRNER IK WHICH LOCALIZ.ATIOK O f

COMBUSTION SURFACE, BACK HEATING, DISTRIBUTION O F HE.\T OVER RADIATISG SCRFACE .&RE .%I,L CONTROLLED

this point and the hearth be long enough. When i t is, the dissipation of heat from the side walls will be large enough t o balance the heat carried down through the solid material surrounding the holes, and the dimensions shown in the sketch have proved t o be adequate for this purpose.

F I R S T P R A C T I C A L S T O V E B U R N E R

I n Figs. 1 7 and 18 are shown two photographic views of the first reasonably practical burner for stove ser\.-ice operating on surface combustion principles-but still imperfect in some respects. To uniformly heat the top bed of loose grain of white alundum or broken fused silica, half an inch thick, thirty-four holes were molded, as shown. by wires inch in diameter and spaced in three rings. This arrangement gave a quite uniform top heat and was operative a t low gas and mixture pressures for an indefinite time, for although the holes of the central group were often red hot for

810 THE' JOUR,V,4L OF I N D U S T X I . I L A N D E Y G I N E E R I N G CHE*MISTKY Yol. 5, No. IO

over a n inch below the hearth the hcat dissipation from the sides was always sufficient to prevent the

Fios. 17 AND I~-FIRBTSUCEESSPUL CooxSrovzTo~ BURNHR. So~ro TUPS

lower end of the holes ever reaching the ignition temperature. The proportions used were selected to give consumption somewhere near those of the standard top burners of ordinary gas cook stoves, tha t is, a t

which is equivalent t o about loo cu. f t . of mixture. Naturally, variations in bed thickness, size of grain or even kind of grain, wi11 cause changes of resistance, the latter due t o the variations in the voids, corresponding to different characteristic fractures. The characteristic consumption tests a t various low pressures are given in curve form, Fig. 3 9 , showing t h a t a t 2 . 5

inches water pressure this stove burner consumed about 1 7 and I j'/z cu. i t . of gas with thc silica and white alundum bcds, respectively, the standard of 15 cu. i t . per hour being equivalent t o 150 cu. f t . per hour per sq. ft. By varying the diameter length and spacing of holes, keeping the general arrangement here indicated, practically any desired rate is obtainable though the general design is the result of the search for a construction suitable for a properly distributed low rate of combustion.

ADAPTATIONS OF TYPE BURNER

To illustrate the ease of adaptability of this general type construction of surface combustion burner to any desired rate over any surface, three other burners are shown, one each designed for a muffle furnace, a crucible furnace and a steam boiler fire, respectively.

The first of these, intended for placing below an assay muffle, is shown in sketch form in Fig. 20 and photographically in Figs. 2 1 and 22, on which latter the chalk line indicates the beginning of the feed holes. This burner had a rectangular hearth, 6 inches by IO

inches, supplied b y 96 boles, molded in alundum cemcnt in a cast iron casing and a t 3 inches of water mixture pressure consumed 1 2 0 cu. f t . of gas per 6 0 sq. inches, equivalent t o 288 cu. ft. per hour per sq. it.

4 still higher rate over an annular hearth was obtained in the small crucible furnace burner of the same general t y p e as s h o r n in Fig. 23, the combustion zone lying between the side insulation and a central cylindrical muffle used t o keep the crucible chamber clear hut not a t all essential t o the operation, though i t helps t o keep the crucible bottom cooler than i t would be otherwise. I n this burncr 2 0 holes, each 0.14 inch in diameter, were distributed so as t u be

plo. 2 0 - ~ ~ ~ ~ r ponM OF n4U~~L.: F ~ ~ ~ A c ~ : B ~ ~ ~ B H , SOL^ TYPE

mixture pressures of z l / * inches water the consumption was intended to be about r g cu. f t . of gas per hour,

on the surface of a cone supplying a hearth lying between two circles, 4'/2 inches and za/6 inches in di-

Oct.. 1913 T B E J O C R X A L O F I N D C S T R I A L

amcter, t he a n a of which is, therefore, I j . 9 - j . q = I O

sq. in. The combustion mas 40 and 47 cu. It. per hour

FIG. ZI-Top Vmw os SOL^ TUPS os MUPI'LB FVRNACB B U R N E I AXLI

WiiiTi: A L O N D ~ ~ ~ M LoosB G ~ A N U L A R BBD

at 2 . j inches and 3.0 inches, respectively, which are equivalent t o 576 and 677 cu. f t . per hour per sq. it. of hearth.

An esample of how this type construction may be

i XIO. 2 2 4 r o u . vmw OP S u ~ r n Tvps MUYYLZ POPINAEG BVRNBR; BE^ IN

PLncE

adapted t o extcnsion of hot surface or hearth over large areas without cxccssive length o i holes in the solid group is shown in the. photograph, Fig. 24, where

Fm 2 1 - C o x s ~ n o c ~ 7 u ~ OP SMALL, S U L ~ T V P ~ MUFPLE FURNACT

three sections of wcdgc form, each I x 3 It. , make together a hearth of nine sq. it. This burner was

.1 N D E X G I RI S G C H E Y I S T R Y 811

tried out in the fire box of the Baldwin locomotive in the Mechanical Engineering Laboratory a t Columbia University as t o its capacity and the proper design

Fro. 24--v1zw CY THRBB Lanos SOLZO TYPE BURNBRS. POKMINO A N ACT. IVE Ben. 3 Fr. SQ., CAPIBLB OP RVnrrmc 20,000 Cu. PI., OP

G A S P E R Hovn

of f a n blast and pipe connection. Mixtures were supplied through a twelve-inch header t o three five-inch hranchcs, one t o each burner, and gas supplied through a six-inch main from the street was measured in a Wylie proportional meter, loaned for the purpose by the Equitable 'Meter Co. At IO inches water pressure the ratc would he 20,000 cu. it. of gas per hour.

P R O B L E X O F FEED II01,ES

While the molding of groups oi feed holes in tapered hanks, long enough t o keep the mixture chamber from reaching the ignition temperature, is a periectly feasible mechanical construction and capable of insuring indefinitely long operat,ion at any desired rate of combustion per sq. i t . of hearth, it has certain disadvantages t h a t warrant a search for an improvement t h a t will remove them. I n the first place i t is somewhat troublesome to mold the plastic material between casing and wires so as to secure a solid mass t h a t will no t crack on subsequent unequal heating, and when large these sections are very heavy and easily broken by shocks tha t may come when setting them in place. Thcy are, moreover, neccssarily high, too high for many locations, such as a cook stove or boiler furnace if present proportions oi frames and settings are t o be maintained. These with some others of minor value are sufficient objections t o warrant a search f0r.a better form though not serious enough t o cause abandon- ment ai this form if no better were available.

Fortunately a better form, and one tha t is completely satisfactory in all respects, has been found and i t naturally follows i rom a little analytical reasoning. 'The holes through which the mixture is t o be supplied must be cooled, and the cooling must he effective enough t o keep the supply end above the temperature of ignition, bu t no more cooling than will do this is necessary or of value. I n addition, t he holes must be spaced a t the hearth for proper distribution of heat

8 1 2 1 L . O F I S D L ~ . S T R I . I L

and thc ilcsircd rate of combustion pcr s q . Et. of hcarth. Banking them in groiips. as has heen sliown, will sccomplish this, but it is clear t ha t complctc or partial structural independence of the holes by forming them in tiiljes so each m:ty he cooled independent of t h e other or evcn forming t h e m in ribs with independent cooling in at lcast t w o directions, is better and more cffcctivc than ii form in which the heat to hc dissipated irom an inner holc must lie ronductcil past ;m outer

doping h m t 01 its own. PK.ACTlChl . H I Y R E R S

Accordingly, this p inc ip lc h a s l xcn incorporatcd, and hurners constructed t h a t sccm, in every way. to

C~ .-.-: ._

i d i i 1

i

PIG A-..T"P V I E W "Y securn F"KM O F COOX STUVI: RIIKABR; C*S119<7 UEWOR>I DKII.I.LN,; .AND Fiar: Bnrcs I .miro

sketch, Pig. 2 j , and photographically in Figs. 26 and 2 ; .

the casting and fire brick lining separate, and Fig. 28. the lining secured in place a f te r drilling the feed holes

I

in each of t h e separately formed cast posts. To insert the lining, wires are inserted in the drilled holes

I 'VD E Y G I Y E E K I . V G C H E M I S 1 K I ~ Vol. j. S o . I O

while pressing the lining in place, the cement forming a perfcct hole in the hearth, csnctly rcgistcring with

I I

- . .. .

P.0. 24-sBcuvD Fonr 0" coor SlUVB IlonnsR CODIPLLlli A N " Knno" sou OPExnrlon

case crushed and sized fusecl silica. This burner has 3.j hulcs each inch in dinmeter hu t othcr diameters

P,,>. i"---R*iri 0 s C"Mn"si*"N A T VAHIOI'S P H I S S U X E C ox C l a C l r L l R H f i a r ~ n 01. Coox STOVE TOP B ~ R N E ~ . 4"<" IN DI.~MBIEK WLTX

33 HrX.Es. BACH 7/64" IN IhAVGTCR "OIL I.OW&.

Cunvs. 3'32" so* IIPPW CilriVB

have been used, as wcll as other matcrial for the l m l of all sorts of thicknesses and while t h e consumption

oct. . lQ13 T I I E J O [ . ' K N A L O F I.VDI'STRI.IL

and top temperature vary with each. practically cvery- thing works ell and any rcsult desired by the designer is obtainahlc merely liy selection of suitable dimensions. In Fig. 30 are shown a pair of characteristic consump-

I .

Fxc. i I - R ~ u i ~ x r ROOM IISATEK EQUI~PED w l i i i SZCONU Puuu OF Tueil- I.AR 'rYPs nnnNEn. S ~ T IN O S ~ C ~ ~ ~ ~ . ~ n s , . ~ , ~ T ~ N ~ ~ ~ ~ vuLcaa

ITBITBR. I,oosx Bsn H s ~ o IN PLACE BY W ~ X E X E S H h-rCtIllOMe SClEBN

tion curves. Burners of this class have been thoroughly tested under both lahoratory and kitchen service conditions, separately and as incorporated in complete cook stoves, having in addition inverted broiler burners with 60 holes, ' / , 6 inch diameter in two rows and a long narrow rectangular hearth with loose granulated material held in place by wire mesh screens. Three of these placed in the top of the Vulcan hotel broiler shell gave consumptions of 38, 43 and j o cu. f t . of gas per hour a t pressures of I inch, i l / z inches and 2 inches, respectively, all making a bright broiling heat t h a t gives excellent service

P I G .?I---RADIANI noor I~EATI:A n u n n e ~ ON BASS W I T H U U T m u OR SIICLI.

TkLis screen construction is wcll shown by t h e iiius- trations of a curved face room-heater burncr, designcd

to fit in a Vulcnn heater shell in the photographs. Figs. S I . 3 2 and 3 3 . Thc first shon~s the new heater complete with an automatic proportioning mixing vaive. t o he rlescrihed later, while t h e latter two show the burner without i ts bed. as set on the hnsc ani1 a separate side riew. This linrner has a rate of 14, r6'.:? and r7'..lr cu. i t . per hour at 1 inch, I';, inches and I ' inches pressnre, respectively, on B hearth X 1 i r x ,j inches = 42'i.l sq. in. supplied with j z holes. l V i inch diameter. and in operation gives as strong a radiation of heat as is consistent with non-blistering of varnish on chair lcgs or ignition of carpet. While the refractory inaterial is brightly hot, the h'o. 18 B. & S, four per inch wire of the screen does not seem t o be above a faint dull red, and at this temperature all the alloys used in electric heater elements intended for operation a t bright red heats up t o 2000' F., without scrious oxidation, will give good service in this location as will also common cast iron grids. Burners of any shape or size suitable for all sorts of household and hotel cooking and heating service can lie produced along these lines to work in any position by the ordinary processes of design. involving n o more complex menta! processes than are required t o seiect suitable materials and dimensions with the nid of established da ta or obtainable by simple experiment in any lahoratory.

This same principle of supplying mixturc thi-ouyh

PI,;. 3.3-SioZ Vmw OB Taemaa BURXSK OF R.AUIAIUT Roo= HSATPK

more or less independently cooled feed holes formed in metal walls and discharging into registering holcs in refractory hearths, has been applied t o higher rate appliances than those for domestic service such as the muffle and crucible service, and one each of these is shown b y the photographs Figs. 34 to 36 both with feed holes driiled in ribs. which alloa~s the same casting t o be used for a v ide range of capacities by changing drill sizes and center distances. The assay muffle is illustrated in Fig. 34. showing the two ribs carrying- the fecd holcs with a few separate tubes placcd at the side and front t o develop cnougb extra heat t o counteract the open front cnd muffle cooling, and equalize thc temperature throughout the muffle. At thc left-hand side is shown a single orifice sampling burner with open bed, consuming two feet of gas per hour and used to adjust the mixture proportions, slight changes in which are readily recognizable hy the appearance of this small open fire. The main hurner is operated normally at 116 cti . f t . p e r hoiir and with the ahove arrangement always gives it very

814 T H E JO17RS.1L O F J 'VDL 'STRIAL A N D E.VGINEERISG C B E M I S T R Y Yo1 j, Lo 10

good thermal gradient, as shown by the following figures:

Txsnaml. G R A ~ I B N ~ IX MUFFLB WITB FLU^ AT Fnom Door closed Door open

Top distance from 1.0 draft Good draft front of muffle Temp. " P. Temp. e F.

2" 1760 1600 4" I870 1742 6" 19II0 1820 8" 1910 1852 IO" 1900 1 I340

Back 12" 1900 I 820

The really significant thing hcrc. hovever. is the easc with which any gradient desired can be obtained by changing the quantity of mixture supplied a t differ-

m;. 3 4 - - . ~ ~ , , ~ i view "Y M ~ ~ ) F , . B P ~ I ~ ~ E B . stilowixc nres, C A R R Y ~ N ~

TU,! MIXTGII: P 6 ~ o HOLIS

ent points, entirely independent of flue location. The crucible furnace shown in Figs. 3 j and 3 6 is fed with two rows of holes ririllcd in a series of radial ribs and discharging in the annular space betwcen insulation and mufie as can he seen in the first illustration with the top removed, which view also shows n sampling burner. The first view shows the side nppcarance and bottom mixture chamher from which thc f e d ribs run upward t o the liase, all cast in one piece. The mixture liolc~s discharge into a hearth of alundum cement on which rests the cylindrical alundum mume and after one run a t ahout 10 inches water pressure, both muffle and hearth softened as shown by Fig. 1 7 on which t h e pcrmanent shrinkaae cracks clearly

adhered t o both muffle and hearth. By a change of alundum bonding mixture both of these parts may be made more refractory, sufficiently so for a considerable

Frc:. 3 S S r o l i v i ~ w OF C K U C ~ ~ L L F r m ~ c r : S u r ~ r ~ e o DY HOL% Dan.r,~o IN namai. RIBS ~ E I W ~ E N MIXTORS CXAMBBR AND FURNACE BOX.

M I X ~ U R E S A M I L ~ X D BURNLR I s SHOWN AT Trle LBFr

quantity of very high temperature work. I t is not difficult, however, t o consume, in such a furnace, enough gas to produce temperatures tha t w-ill soften any available refractory, bu t by simply changing mixture pressures i t is easy to maintain practically any temperature desired. If for any special purpose i t should he necessary t o maintain a secondary hot zone a t any particular point, then separate mixture supplies can be led t o tha t point and burn without interference with the products of combustion passing from another combustion zone, a condition peculiar t o surface combustion, to which the well known "smothering"of common burner flames is unknown.

111. CONTROL AND ADJUSTMENT OF K A D I A T I K G

SURFACE

By the application of the principles of construction so fa r developed, it is possible t o make a surface com-

Sic;. J G T o P V ~ H W OF CEUEIBLE P u n ~ ~ c r : wmii COVEPI RBMOW?D

bustion burner for any desired rate of combustion per sa. i t . of liearth tha t will oueratc for indefinite Deriorls _ . -

show as do also some pieces of white alundum t h a t of time without disturbance or adjustment, bu t for

Oct., 19'3 T H E 10 UR.ZTA L OF I.11 I1 I ' S T R I A L

commercial apparatus even more than this is needed. One additional property that must be present is a suitable range of control of the rate, as few, if any, burners are operated in practice a t a constant rate of combustion whether used for domestic or industrial purposes. It must be possible, therefore, t o adjust the consumption of any burner from a minimum low, t o a maximum high rate and t o operate safely and surely a t any one, no matter what the previous adjustment or the time of operation. Furthermore, some classes of scrvicc rcquirc a very prompt change in the radiating bed temperature after adjustment, which includes a very quick initial heating after lighting cold, while other classes of service require a slow change after adjustmcnt or a sluggish response t o changes. With any given range of adjnstmcnt of rates for safe working,

Fro. 3 7 4 ~ ~ ~ Vrsw OP CP.IL'EIBL& FURNACE MVPPLF AND HEASTR. SHOW- TNC THE EPPacr OP SOPTBNINO 0 s MuPSLE B Y HEAT. MATBRZAL

oP BOTS. ALONDOM M ~ x T u R E

there may be associated the requirement of a prompt or a sluggish response after a change, bu t fortunately in no one burner is i t ever necessary or even desirable t o provide for more than a limited range of adjustment associated with either prompt heating or sluggish, never both. These conditions can be met in the sort of burners described sufficiently well t o mect practically all service requirements consistent with maintaining the advantages of snrface combustion as a process.

RANGE OF COKTROL OF RATE OF CO.VBUSTION

The range of control of rate of combustion depends primarily on t h c size of thc feed holes and on the effectiveness of their cooling, and i t is easily possible to make lioles small enough in separate tubes or ribs t o absolutely prevent back flash even when the mixture is turned off and the velocity of flow reduced t o zero. By imposing sufiicient mixture pressure on

A N D ENGI.VEiIKI.VC C I I E X I S T R Y 81 5

enough of these holes per square inch of hearth, any maximum rate of combustion may easily he associated with this minimum rate of zero. Certain practical considerations such as the extra cost of farming small over large feed holes and the increased possibilities of stoppages in them make i t desirable t o use larger holes than correspond t o a minimum rate of zero when tha t is not necessary, and i t has been found tha t holes 1 / ~ ~ inch in diameter are small enough for safe operation a t rates as low as most service requires. Holes of this size in cook stove top burners, broilers and room heaters, can be safely operated a t two- or three-tenths of an inch pressure, giving thereby a t least as great a range of adjustment as is possible with Bunsen burners of the ordinary form without back flash. Such holes seem t o keep free of stoppages and me cheap t o form if not much over one inch in length. Apparatus intended normally for much higher rates than the domestic class w-auld naturally be given larger holes, and, therefore, have a higher minininm operating pressure, but the range desired even with this higher minimum is easily obtained b y using a correspondingly higher maximum. Thus while the maximum pressure available for operating domestic apparatus without gas boosting, which is out of the question here, is a little less than tha t of the street main, say 3 inches of water, t h a t for an industrial fnr- nace for which gas boosting is not objectionable may be made even with small fans as high as 12 inches and with positive blowers several pounds. Assuming the holes in the latter case t o be such t h a t the minimum safe pressure is I inch, there is a pressure range of 1 2 : I for adjustment of rate. Correspondingly, if the domestic apparatus has a minimum safe pressure of 0.2 inch, then i t has an adjusting pressure range of i j : I . I t must not be understood tha t these are limiting figures, for they are not, being offered merely as examples, since the real range may be made whatever is desired, a thing thought to be impossible in the early days of surface combustion development.

TIME OF HEATING

The time jf heating is likewise within control though not through so wide a range; i . e . , by adopting suitable constructions the burner may be made t o heat up from the initial cold t o the normal working s ta te in a short or long time, and will be correspondingly quick or sluggish in response t o pressure rate adjustments while in operation. Sluggishness is obtained by nsing B thick bed of great mass and, therefore, large heat storage capacity, while sensitiveness follows the use of beds or radiating surfaces as light as possible. I t is clear that , taking for example the cook stove top burner, a bed only inch thick will Beat quicker and respond more positively t o valvc adjustments than a bed I inch thick, other things being equal, and the t ime of heating the latter would approximate four times tha t of the former if there were no interfering influences. These influences are, however, present and are somewhat difficult to understand a t first, bu t clcar enough after study.

I n the first place, a reduction in the thickness of bed may leave it so thin tha t i t is no longer effective

816 T H E JO17R-VAL O F I ; V D C S T R I d L A N D E N G I Y E E R I N G CHEAIfISTRY T'ol. j, SO. I O

in reducing mixture velocity a t t he hearth orifices, so t h a t mixture will blow off at various points where t h e grain does not properly baffle a jet. Th? escape of this mixture actually reduces the heat developed in the bed, delays i t s heating and so counteracts t he expected gain from bed mass reduction; a t t he same time i t causes a smell from the escape of t he unburned gas which, of course, ceases once t h e bed is heated even when the baffling is very poor. This is one of t he difficulties encountered in the effort t o reduce the initial heating t ime by reduction of bed thickness. I t would seem t h a t if, start ing with grain '/4 inch in diameter a reduction of bed thickness was carried so far as t o cause partial blow off by insufficient baffling, this could be corrected b y a reduction of size of grain and this is so, b u t here again limits are met. I n no case may t h e grain be so small as t o fall into the end of t he tube , bu t with small holes used, even before this occurs two other things are noted: ( I ) A th in layer of fine grain on the flat hearth will not heat uniformly with the hole spacing satisfactory for larger grain; a series of bright spots above each hole with dark areas sepa- rating them will replace t h e uniform color of a thicker and , therefore, more sluggish bed. ( 2 ) The fine grain bed will not heat initially in the same way t h a t t h e coarse one will, for below a certain grain size the combustion locates initially on the outside surface, which heats first, t he combustion zone gradually working back toward t h e orifices, while with a larger grain, offering fewer and wider mixture passages in t h e voids, t he combustion zone locates instantly about t he holes, flashing directly through the bed. I n other words, t he fine grain bed heats from the surface iizward while the coarse heats from the inside outward, and though for two such beds of equal weights of material i t would seem tha t equal times must elapse before each had reached a steady s ta te , as a mat te r of fact t he one t h a t heats first from the outside will be quicker because of the non-escape of unburned mixture while heating.

Q U I C K H E A T I X G T O P B U R N E R S

This reasoning leads naturally to one of t he successful forms of quick heating top burners which becomes incandescent and uniform a t t he normal ra te of combustion in about half a minute, t he ordinary half- inch thickness of large grain bed requiring three t o four minutes, depending on the material. This one is formed by counter-sinking t h e outlet side of t he

inch feed holes t o '/z inch, filling t h e t ape holes with silica of size 10/64-2/64 inch, and also covering the hearth t o a depth of ' / 4 inch. as indicated in the sketch, Fig. 38.

Reduction of mass t o be heated without interference with t h e effectiveness of t h e baffles and by using materials t h a t are permanent is a definite problem for which there are many solutions and no doubt others will appear as long as specific attention is given t o i t , bu t i t is a serious question whether reduction of time below one minute, or even t o this t ime, is really advisable. There is available one type group of construction already explained under localization, and t h a t is t he formed baffle, of which the flat plate mushroom

plug and ball i n tape hole are examples, and over which may be placed the radiant screen, of metal, perforated or woven, a th in layer of loose grain, or a diaphragm of refractory perforated or otherwise formed. to be porous enough t o allow the free passage of gases; all these have been used as well as some others t h a t proved unsatisfactory. I n the latter class fall a layer of picked asbestos fiber which heats as quickly as a n illuminating mantle b u t which has no life; a inch layer of hollow cylinders or short th in silica tubes like straight beads, ' /8-' /4 inch diameter and 3 / 8 inch long, which heats in less t h a n one minute and is permanent

. bu t too expensive; and a thin plate of alundum cement perforated freely with '/8 inch holes, which becomes incandescent in one minute bu t is too fragile. I n the former class fall t he perforated metal non-oxidizing alloy plate over some form of direct baffle, and the molded porous diaphragm directly over t he holes serving as both baffle and radiator, t he former giving a very quick heating bu t not so permanent as the la t te r

w

FIG. 38-hfODIFICATION OF TUBULAR T S P E COOK S T O V E TOP B C R N E R

FOR QUICK INITIAL HRATING

which. however, heats more slowly. For the metaI plate, two materials have been used, Nichrome I1 and Excello. the analyses of which are given below, as determined by Whitaker and Metzger:

ANALYSIS OF P\TON-OXIDIZING ALLOYS (PERCENTAGES)

Nichrome I1 Excello

Sample No. 1 Sample No. 2 Nickel.. , , , , , . , , , . 5 6 . 4 5 9 . 9 Nickel.. . . . . . 8 3 . 4 3 Chromium., , , . , , , . 10.6 10.8 Chromium. . . 14,72 Iron.. . . . . . . . . . . . . 1 . 3 0 Manganese.. . . . . . . 2 . 1 2 . 3 Silica . . . . . . . . Trace

9 9 . 7 9 9 . 2 9 9 . 4 5

3 0 . 6 26 .2 Iron. . . . , . . . , ~ ~ ~

The former has been most used and as applied to the problem of quick heating was in sheet form, 0.032

inch thick, perforated with holes 0.073 inch diameter a n d about one diameter apart . This was cut circular t o fit t he top burner hearth 43/8 inches diameter and heats t o normal in about half a minute when set over baffles of t he mushroom or plate type , and in about ten seconds with ball baffles. Such plates have a fair life if not operated at too high a temperature; about one cu. ft. of gas per hour for each two sq. in. of radiation is a satisfactory rate and gives a good red heat. This construction may, therefore, be adapted as a direct competitor of t h e electric heater, using the same radiant material a t the same temperature, bu t with gas as the source of heat instead of electric current. A series of consumption curves of one of these quick heating nichrome radiating plates, 43//8 inches di-

OCt.. I 9 1 3 T H E J O C RLV*4 L 0 F I iV D US T RI A L 11 -V D E ,\'GI E E RI,VG C H E MI S T R Y S I 7

ameter set over 3 j holes, inch in diameter, in t h e cook stove top burner hearth with mushroom baffles plugs is given in Fig 3 9

- . Lt.-3it.-crb _ L ?i-kll'-&-L+~~-E~ -I- ~ ! ! ~ t - ~ ~ ~ ~ i ~ c ~ - t ~ ~ i i t " f i ~ i '

F I G . 3 9 - R A T E O F COMBUSTION O F TUBULAR TYPE C O O K S T O V E T O P

R U R X E R HAVING 3.5 HOLES, EACH 1/16" I N D I A M E T G R , DISCHARGING U N D E R S Q U A R E S T E M M E D hIUSHROOM P L U W , S H O W N IN F I G . 3 A N D

P R O V I D E D W I T H P E R F O R A T E D x I C H R O M E RADIATING P L A T E

REDL7CTIOT OF H E A T I S G T I M E

I t is a serious question whether the reduction of heating t ime should be carried so fa r , even in domestic appara tus in spite of a demand on t h e par t of users for i t : because of t he corresponding small heat storage. The top burner carrying a half-inch loose bed has so much heat storage t h a t i t will resist interference from t h e spilling on i t of large quantities of water or grease, amounts of which sufficient t o be serious with Bunsen burners have no effect here whatever, b u t t h e more this heat storage is reduced by constructing for quick heating t h e less this resistance. Another advantage of heat storage in t h e incandescent bed is inherent i n all intermittent heating such as t he heating of laundry irons. which may thereby absorb heat for a time faster t h a n it is being generated by drawing on t h e bed storage; these always heat faster on the surface combustion burner t h a n on a common burner of equal gas consumption. One trial showed t h a t with two irons on a surface combustion burner a laun- dress could do as much work as with three irons on the Bunsen, t he former consuming about half t he gas t h a t t h e latter used a n d heating the iron in less time.

I n applying these surface combustion burners a new s tudy of service conditions is necessary as they have some characteristics, no t possessed by the old burners, t h a t permit a n d even demand adaptation of burner design t o service. There will always be some cases where quick heating is impor tan t , e . g. , in toasters; others where heat storage is impor tan t ; t op burners for intermittent service and those tha t are much used and likely t o receive spillings; also others t h a t will demand t h e smooth self-supporting surface of t h e bonded diaphragm, which is operative in all positions and which has properties all i ts own.

I S F L U E S C E O F S E R V I C E COSDITIOPU'S

L-SE O F D I . 4 P H R A G M S

The application of t he bonded diaphragm t o general service conditions is a problem in itself, as t he localization of combustion is not so readily controlled under wide ranges of mixture pressures within t h e limits imposed b y t h e ordinary domestic gas service. A11 such diaphragms except those of very great porosity and , therefore, necessarily thick, will heat from the outside inward and in all, t he heat will work back. t he combustion zone following until back flash takes place under some supply pressures and times of heating. Assuming t h a t general service appara tus will be operated a t a variable ra te and a t such variable pressures, t h e localization combustion by diaphragms is no t t o be depended upon, bu t they will still be useful as baffles and radiators when placed over a hearth, such as described, and which latter insure localization.

Placing a bonded diaphragm over such a hearth as t h a t of the cook stove top burner, i t will, on ignition, heat from the outside first; t h e combustion zone will then work down a n d through i t until a back flash occurs and t h e combustion zone locates between the hearth a n d bot tom face, extending from each hole into t h e diaphragm a little way if i t lies directly on the hearth. As the hea t works back unevenly, due to non-homogeneity of voids and size of grain, there u-ill be at the moment of ignition below i t only one hot spot, all t he rest of t he bot tom and the hearth being cold. Up to this moment the outer surface has been hot and radiating, b u t now the heat will be practically all absorbed by the hearth and lower space until they become incandescent, and in this period the top radiating surface will cool and then slowly recover i ts temperature. after which i t will operate indefinitely without change. The t ime it takes for t he outer surface t o begin to radiate, and for t h e back flash t o occur are both dependent on the relation of porosity to pressure or ra te of supply. bu t t he t ime of recovery after back flash is largely controlled b y diaphragm uniformity. If t he combustion zone works back equally fast all over t he diaphragm t h e bot tom face will be hot all over when bot tom ignition takes place, instead of only one spot, and thus the top temperature will be recovered most rapidly.

A great many small diaphragms of all sorts of grain materials and bonds have been tried and all abandoned in favor of a lundum; these have been prepared b y the Norton Co., t o whom considerable credit is due for t he spirit of cooperation displayed in opening up a new field, absolutely without da t a . Without reciting t h e details. i t may be noted t h a t satisfactory results are obtained with S o r t o n alundum diaphragms fed through the self-cooling mixture holes and set over t h e discharge hearths. These are structurally quite strong if t he bond used is not intended for very high temperatures, in which case they are delicate bu t still useful for some purposes. Two consumption tests of such diaphragms set in t he hearth over 3 j holes, '/I6 inch diameter, t he porous par t being 4 inches diameter a n d 2 inch thick, are given in Fig. 40, comparing a Norton a lundum diaphragm of 90. 4 grain with a n English fire clay diaphragm, which latter fused in

818 T H E JOCR,VAL OF I N D U S T R I A L A N D E X G I N E E R I N G C H E M I S T R Y Vol. j, No. I O

t ime on the hearth side, t he former remaining unchanged. I V . A U X I L I A R Y A P P A R A T U S I n Fig. 41 are given similar results for a series of alun- Any burner or furnace intended for laboratorv

F I G . 4 S R A T E OF COMBIWTlON OF COOK STOVE TOP BURNER HAVING 35 HOLES, EACH 1/16" DIAMETER. DISCHARGING TO BONDED DIA-

PHRAGMS. TOP CURVE DIAPHRAGM OB NORTON ALUNDUM MIX- TURE. BOTTOM CURVE ENGLISH FIRE BRICK MIXTURE

d u m diaphragms j1/8 inches diameter, inch thick, set over t h e same hearth after enlarging. These were of four sizes of grain No. 8, No. 6, No. 4 and No. 2 ,

and three different processes of working. Change

F I G . 4 1 - R A T E OF COMBUSTION OF COOK S T O V E TOP BURNER HAVING 35 HOLES, EACH 1/16' I N DIAMETER, DISCHARGING THROUGH NORTON

ALUNDUM DIAPHRAGMS OF DIFFERENT CONSTRUCTION. E A C H DIA- PHRAGM 51/s" DIAMETER AND 7/16" T H I C K . VARIATION IN

CURVATURE SHO ws TRANSFER OF MAIN RBSISTANCE FROM FEED HOLES TO DIAPHRAGM

of curvature indicates t h a t t he main flow resistance has been transferred from feed holes t o diaphragm.

use where operators have sufficient skill and under- standing of cause and effect, need not be supplied with any special appliances for proportioning air t o gas in the mixture, or for changing the mixture pressure as service conditions require i t and, therefore, any of the many s tandard fans and blowers may be used for compressing air alone or air and gas separately with a hand-adjusted cock in the air and the gas pipes supplying each appliance. This is by no means t h e case, however, with some industrial and with all domestic appliances. Where a n industrial furnace operator is more or less constantly engaged in managing one furnace, he can be taught t o make adjustments of hand values if provided with the ordinary water column water gauge and a small sampling burner, bu t this is entirely out of t he question for the ordinary woman in the home or chef in a hotel, as they either cannot o r will not adjust anything beyond turning the cock t h a t regulates t he heat and i t must be confessed, sometimes not even this. For domestic service, including t h a t of hotels and restaurants, i t may reasonably be expected t h a t they operate a single cock for each burner to regulate t he heat, bu t maintenance of mixture proportions essential t o surface combustion processes. must be automatic and fool-proof. This a t first, like some of the other questions taken up, looked like a difficult problem yet it seems to be working out in a thoroughly satisfactory way.

A C T 0 PI A T I C MIX T C R E P R 0 P 0 R T I 0 SING

Automatic mixture proportioning has been accomplished by two devices which working together carry out the fundamental method; i. e., proportioning by equal pressure drop on a n air and a gas orifice, having areas in t h e desired proportion. Regulation of air pressure t o an equality with t h a t of the gas supply available a t the appliance will bring them t o a controlA

F I G . 4 2 - s E C T I O N A L VIEW OF DOUBLE-PORTED MIXING V A L V E , W I T H AD- JUSTMENT FOR RATIO OF AIR TO GAS

The t ime of back flash and recovery of surface temper- a ture varies considerably b u t in the best t he whole t ime is about five minutes; in some cases it has reached I j and even 2 0 minutes. The t ime of recovery varies f rom 3 t o I 2 minutes, t he better diaphragms requiring the lesser time, all a t pressures of inches of water.

cock having two ports with areas in the desired proportion and by suitably large pipes or small por t openings; these ports determine the flow t o a common mixture outlet, and the proportions will remain constant so long as the principal resistance t o the flow of both

Oct., 1913 T E E J O U R N A L O F I N D U S T R I A L A N D ENGIiVEBRI.VG C N E M l S T R Y 819

is in the cock ports. This is accomplished, first, by sufficiently large supply pipes b u t i t is necessary also t h a t the ports bear a suitable relation t o the burner opening so t h a t while the pressure drop through the burner is large enough t o establish the desired veloci-

tion a t the left. This bas two plugs, keyed together so as t o rotate together, hu t otherwise free; one, carrying the small gas port , is tapered and spring seated for tightness, while the other, carrying the air port, is cylindrical, and has an axial adjustment for length

ties for localizing the combustion zone, there is also some drop of pressure through the ports themselves. In other words, the combined area of the ports in the wide open position must never be so large as t o build up in the mixture chamber a pressure equal t o tha t in the air and gas supply pipes. This does not involve any serious drop through the ports; two- or three-tenths of a n inch of water is ample, and less will do if necessary, b u t the more there is the more definite the proportioning and incidentally the smaller the valve needed. To permit of an adjustment t o changes of gas quality onc port may be made of fixed size with the other adjustable, bu t in opening and closing both

of effective opening t o adjust for gas quality. Both plug ports are aligned with the casing ports so as to absolutely open and close together. With this cock, the user regulates the burner exactly as has been the case with common burners, controlled by single ported cocks.

CONTROL op PRESSURES

I t is necessary, however, ior the satisfactory working of any form of double ported cock, intended t o maintain constant proportions of air and gas while controlling the total quantity of flow from zero to a maximum, tha t the air and gas approach at equal pressures

must vary in constant ratio. I t is not expected t h a t users will make this quality adjustment, bu t rather the gas men who make installations and, of course, in any one district a setting will be found good for the average air requirement of the sort of gas supplied, the variation in wliich is never large enough to seriously interfere with the operation on this mean setting. Wherever the gas is uniform or substantially so, in air requirements. this gas quality adjustment can

Pia. ~~----S~CTIONAL VlEW OF AIR P x H B S C ~ S R S G U L A T ~ FOR MIIINIAININE Ala PRHSSURH EQUAL TO THAT on GAS

be eliminated and the ports fixed at the proper value, once for all. One successful cock of this sort is sketched in Fig. 42 and shown photographically in Fig. 4 3 , complete on the right and in various stages of dissec-

PIG. +S--VILW us RBCDL~~TVR COMPI.BTB: CONTROL D A M P S ~ HELD OPBN BY * RN,l’*

so tha t the drop in pressurc to the common mixture pressurc be cqual; this may be accomplished by regu- lating the air pressure t o tha t of Lhe gas, or conversely. The former has been used as most satisfactory. and in accordance with i t . the air delivered by tlie fan i s passed through a diaphragm regulator carrying a damper valve in the main air passage. This damper is rotated by a crank and connecting rod, fixed t o tlie meter leather diaphragm, the weight of which is bal- anced by a spring or dead weight, if horizontal, but is

8 2 0 7 ' t lE 1 0 C ~ K . > ~ . I L O F I .VDl-STRI . iL

t o he neglected if operntrd in a vertical plane. The dclivcred air piessurc is caused t o act on the bottom of the diaphragm vhile the local gas pressure ac1.s nhave; both sides, hcing dead ends. tha t is, not swept b y any current, there islittic. h d e n c y t o dry theleather. All changes o i gas pressure vary the diaphragm load and move thc damper t o the position for re-equaliz:i~

I

P , ~ . + 6 - . ~ , ~ rRIISSU,t~ nmrl.lToK P A R ~ ~ Y ASS^:,^^^^^, sxowruo .,v rliE C E H T ~ ~ CONYBCTINU Roo AN$> Cnhsx *OR ACIIT&TIPC THB

D A M P S R SY in4 Di'rPaarr,M M",-E\ILNT

tion promptly and positively so tha t no matter how local gas pressures may vary. the air automatically follows and the mixture quality remains what i t shoiilii be. The construction of this regulator is shoivn hy

. I SU L.VCI.VEERI.VG <~Irfi.viI.sTRy TOI. 5, SO.

tion of these surface comhustion appliances but the pressure nced he no mor? than is necessary for good regulator action, for which ahout I inch mater drop is de- sirablc. Air pressures o i this ordcr of magnitude are easily ohtsinnhie hy fans ani1 arc preferably. though not necessarily. clcctricirlly driven. If thc full available gas pressure is used, and this be three inchcs, then t h e fan should deliver a t four inches of vater, bu t very satisfactory operation can hc secured at mis- tu re pressures as low as 1.j inches, for full capacity. which u~oi~lii rcquire l css than t w o inches of gas pressure and less than three inches fan dclivery. At such pressures the powcr requirements of the inns lor ordinary appliances are very small, as may h e easily calculated. For quick refcrencc, the results of such a calcuiation are gi\.en in chart in Fig. 47 for a wick range oi capacities and pressures, t o which is added a scale for rending motor horsc power for any inn cffi- ciency and e1ectric:il inpiit horse bower for any electric motor efficiency. For small sets the fan eficiency will probably he iictween 3 0 and j o per c e n t , while

8 A

A

8 2

F,o. 4i---Cs*nT SlilowiNr; TI34 aunsr Powen AND C"R.$NT CORST~MPTION FOX MUTOR-DRIYBN FANS wvrc VARLOUS

DR~,~IVBRV PRaSSUs6S. Q u ~ N n r r s s OF A m Pfin PrlmlrTE, FAN AND MOTOR E F I - I C I ~ N F Y

With these auxiliary appliances, i t is

,...-... ., .~,... ..... combustion system for practical service in either onc of three ways: ( I ) Each appli-

ante with its own fan and regulator and so completely independent except for connection to gas and electric service mains. (2) A central fan with dis- trihuting air pipes t o each appliance with its own regu-

Air under pressure is always necessary for the opera- lator. ( 3 ) Camhinations of appliances in groups.

the sketch, Fig. 44, and the two photographs, Fig. 45 showing one size of regulator complete. aIld Fig. 46 showing i t partly dissected

AIR P R E S S U R E h - E C E S S I R Y

OCt. . 1913 T H E J O l . R 8 A L OF l S D l ~ S l R I , I I

I t is compnrntiveiy casy t o see how central or group systems would work out for large installations suitable, for cxampie, for hotel kitchens, hu t it is a matter of some interest to cramine the independent domestic unit.

FIRST S C R P A C E C O \ 1 H l I S T I 0 4 K A S G E

The first of these ever constructcd was made by converting one oi the standard side cahinct ranges now on the market. using mixing valves t h a t had been used elsewhcre and a fan taken from n vacuum cleaner, in the absence of one of suitable capacity as t o volume and pressure. The fan used was t h a t of the Regina cleaner in onc casing with a G. E. motor and had nu air volume capacity about four times what was needed, delivering at ahout I I inches of water where four ox five wonid havc been sufficient: i t consumed 00 watts with all hurncrs in operation. These things are stated l o explain the somewhat crude appearance of some parts of this, the first self-contained surface cornbustion range ever constructcd a n d pu t into service in a home kitchen. I n Figs. 48 and 49 a11 parts are ciearly shown,

nS. ~ B - D ~ ~ G O N A L TO? VIBW OF F ~ ~ ~ T noMeSnc RANGE; SOL^ CAST IION Top, WITH Foua Tor HVRNEKS; R B P K ~ C T D R Y IN PSAC&

BU'C B u ~ ~ e n Car" RBMUVBD

except the check valve between f an delivery and regulator t o prevent g3s escape from the ian , should a burner be turned on when the fan is not in operation; this was not at first provided. I t will be noted tha t the miring valves or double ported cocks are fitted with graduates t o jiuide the operator in adjustment, as t h e absencc of visible flame, the ordinary guide, here dcmands a substitute. All auxiliary or secondary nir openings into the ovcn are closed as now unncces- sary and if left tending t o rcduce efficicncy by the air so drawn in and discharged hot to the Hue. For the same reason the chimney vent is made smaiier. The oven Hue linings were not changed hut should havc been

. l . V l l E,VGISEEKISL' ( ' I I E M I S T R Y 8 2 1

t o provide for the flow oi a smaller volume oi gases a t higher temperatures with the siirface combustion hurners t h a n with the original Bunsen burners.

Flu. 49-Fmm VISI OF FMST COMPLBTB DOMFISTTC RANGE, WTX O V ~ noons i m N . B ~ O I L E R *ND ovrru B ~ R N E ~ . A=I V I S ~ S L ~ . nncx nrrso nm.mvLI . , MLXTURE S I I P P I . ~ TO HOLBS nxrLmn IN n m S B T W F I ~ MIITORE C I I A M ~ E R A N D HBARIII. BVRNSRC * R E SYSPHNDBD PROY Fnonr AND BACK WALLS. T i r ~ C r o s i ~ o OP T H ~ OPBNINGS BO* SECONDARY A m 1s ALSO VISIBLION ?~&I.owGR Dooa, NO SECONDARY Ala Bzrnli N S E D I ~ ron

Ti l s SUKIACI: COM805ri"N P R O C I S S

V. EFFICIEA-CY OF SURFACE C O X B U S T I O N APPLIANCES C O X P A R E D WITH BUNSEN

As in other cases. i t is here found t h a t the absolute and relative efficiency of surface combustion and Bunsen burner appliances is no t a constant thing but varies with conditions and, therefore, any discussion must be more or less limited in scopc t o give a correct impression. For this reason the figures obtained from tests will he confined t o domestic appliances operating on city gas, such as cook stove top burners, broilers. oven, room and water heaters, all made by removing Bunsen burners from standard appiiances and substituting the new ones. Top hurners were tested by taking the consumption and time to raise a weighed body of water t o zoo' F., from an initial low temperature; the same or identical vesseis aiways open were used in all comparative tests.

From a large mass of tests conducted with various bed materials and thicknesses, different sizes, shapcs and materials of vessels; different distances between bed or fiame and the vcssei, a t all sorts of gas or mix- tu re pressures a t various positions of the C O C ~ S a n d with rarious stove. top or burner ring construction, thc figures tl~emselves i n r y so much as t o be a t first perplcxing t o onc seeking some definite, general conclw sion on the relative value of the t w o systcms. Careful analysis, however. reveals some order from which con- clusions foilow relating results to conditions. The

a 2 2 T H E J O 1- RS.4 L O F I ,V D L-S T RI A L

first of these is t ha t with identical conditions or as nearly so as the differences in the systems permit, t he surface combustion burners always do a n equal amount of heating of t h e water with less gas than the Bunsen b u t how much less seems t o be dependent on the construction of burner, stove frame and water vessel. Perhaps the most important of these external conditions is t he vessel itself, and however queer this may a t first seem, an explanation is available. If t he vessel be large, of good conducting material, like copper or aluminum, and containing a large amount of cold water, i t would approximate the heat absorber of a gas calorimeter and could take up a very large amount of t he heat being generated a t a given rate. On the other hand , if t h e vessel has a small heat-absorbing surface in proportion to the ra te of combustion and be made of poor conducting material then there is less chance of heat being taken up by the water, other things being equal. Size and material of vessel placed over a burner developing heat at a given ra te do, therefore, determine, t o some extent, t he rate of heat absorption, bu t also t h e ra te of heat dissipation from its sides, both the submerged and exposed part , and the water surface. This last factor is also more or less influenced b y t h e extent t o which the vessel is filled a n d the air drafts blowing on i t . It thus appears t h a t a great deal depends on the vessel and i ts filling, bu t as many detailed figures are always confusing the results of twovessels only are given in this connection.

One of these is a cylindrical tinned iron sauce pan, about 8 inches in diameter and of equal height and the other a smaller tapered pressed steel enameled sauce pan, about 6'/2 inches in diameter, both tested on New York City gas, over burners consuming ~j cu. ft. per hour, t he Bunsen burner being of t he Crane construction and the surface combustion having a 4 3 / ~ inch hearth with 2'J/64 t o '4 /64 inch material, '/z inch deep and covered a t t he edges by a cast iron ring with vents between it and the vessel l / g inch wide. For equal amounts of heat taken up by t h e water t he relative consumption was as follows:

Bunsen burner consumption = 1.44 X surface combustion burner consumption with large iron po t ; 1.66 X surface combustion burner consumption with smaller enameled pot. These two burners tu rned down t o a simmer, using the enameled pot a n d keeping the water just below the boiling point as nearly as could be judged by the eye, t he consumptions were: Bunsen 5.2 cu. ft. per hour and surface 2.42 cu. f t . , giving t h e ratio of 2.15.

The average of t h e two other series with enameled vessels on Westchester, N. Y. gas, a thick silica bed surface combustion stove, grid top, I j cu. f t . per hour, with a G. G. A. Co. Bunsen burner, gave consumption ratios of 1.71 and 1.75. A different surface combustion burner, t ha t designed for quick metal heating becoming normal in less t h a n one minute, as compared with three or four minutes for the above burners, and illustrated in Fig. 38, gave a ratio on N. Y . C. gas of 1.80. This last was also fitted with a grid top, the enameled vessel being held about inch above the bed.

A N D E,?rGIA7EERING C H E M I S T R Y v01. 5 , NO. IO

These are high values a n d indicate a large fuel consumption as may be seen from the following figures: Assuming a Bunsen burner t o be consuming I j cu. f t . per hour, then t o give the same heating effect in a cooking vessel, t he surface combustion burner for consumption ratios of 1.8, 1.7, 1.6, 1.5, 1.4, and 1.3 would require 8.3, 8.8, 9.4, 10.0, 10.7 and 11.5 cu. f t . per hour, saving 6.7, 6.2, 5.6, j.0, 4.3 and 3. j cu. f t . gas per hour while doing the same work. The high ratio above reported as actually attained may not be under favorable conditions and low values have frequently been found; i n one case lower efficiency with a certain surface combustion burner and top construction than with the Bunsen; bu t this fact pointed out the mistake in construction which, when corrected, resulted in high ratios. As has already been pointed out, t he size of vessel is important bu t so also is t he heat dissipation from the top. I n one case a solid stove top was used with circular openings t o take the burners and quite low efficiencies were obtained, as low as 1.2, bu t i t shortly appeared t h a t most of t he hot gases were escaping along the underside of t he cold top instead of rising around the vessel, due t o bad construction. A simple substitution of a grid t h a t allowed t h e gases to rise with no other change raised the ratio from 1.2 t o 1.68.

GAS SAVED B Y N E W PROCESS

It may easily happen t h a t service conditions will arise t h a t will demand constructions t h a t are not highly efficient as, for example, cold cast iron tops for intermittent service, in t he interest of cleanliness, so tha t a commercial stove construction will probably be a compromise. The best judgment t h a t can now be made is tha t , taking into consideration all these conditions, there may be expected a consumption ratio of about I. jj Bunsen t o flameless, or 0.65 flameless t o Bunsen, which indicates a saving of about 35 per cent of what is now being used though i t is possible according t o actual tests t ha t i t may be as high as 45 per cent.

E F F E C T OF EXCESS AIR

It has previously been stated tha t t he expected savings from the surface combustion process were t o be derived first from the elimination of t he excess air now required of all Bunsen burners, and second by the superior penetrating power of radiant heat in warm- ing bodies from the fire. The influence of these two together is well shown in all t he preceding relative consumptions b u t perhaps nowhere quite so well as in the case of the simmering test where two fifteen- foot burners were turned down t o keep the water just a t the boil for about one hour, which raised the consumption ratio from 1.66 t o 2.1 j. This is a splendid demonstration because in the turned down condition very much more cold air rises with the hot flame gases t o the pot bo t tom in Bunsen burners t han when they are on full while there is no change, whatever, in this respect with the surface combustion burner. The difference above is, moreover, largely due to excess air because a t the turned down condition the surface combustion burner bed was only faintly radiant.

OCt., I 9 1 3 T H E J O U R N A L O F I S D U S T R I A L A S D E , V G I S E E R I S G CHEMISTHI' 8 2 3

The minimum effect of this excess air influence is perhaps best obtainable f rom a comparison of two water heater tests made on identical coils of t he Vul- can heater, one as constructed with s tandard Bunsen burners and the other with a surface combustion burner substi tuted, bu t without other change. Both were operated t o heat the water in a s tandard forty- gallon copper tank , equipped with thermometry and the results are uncorrected for radiation from it or pipes. On an assumed calorific value for the gas, the Bunsen burner gave an efficiency of 70.4 per cent, while the surface combustion burner gave 8 2 . 3 per cent, a ratio of 1.17, the flue gas temperature of t he Bunsen being 248' F. against 309' F. or 61' higher, for the surface with i ts excess of I 2.3 per cent efficiency. Of course, a properly constructed surface combustion heater can easily be made to yield as close t o IOO per cent efficiency as may be desired; these figures are given not t o show what may be expected of such, bu t purely as a measure of the excess air influence when it is reduced t o i ts lowest value in the Bunsen equip- ment ; this condition is entirely eliminated by surface combustion.

Another measure of excess air influence is given by the figures for a comparative top oven test in which an upper oven was heated by burners below, intended primarily for a broiler oven below, Figs. 48 and 49. -4s originally equipped with Bunsen burners this

' top oven required 19 cu. ft. of gas per hour t o maintain a t its center a temperature of 480' F., as against 13. j cu. ft., a ratio of 1.4, when surface combustion burners were substi tuted, with no other change t h a n the closing of the secondary air openings, linings which should have been changed being left in place.

VALUE O F RADIANT HEAT

The direct value of t he radiant heat developed by surface combustion burners is shown most clearly in test of such apparatus as a broiler, or a radiant room heater, Fig. 3 I. Tests have been made in such broilers by taking the consumption for t he proper cooking of steak and chops, both thick and thin, chicken, squab, fish, lobster, and toast , under the direction of competent chefs, and comparing with t h a t required for t he same carefully measured and weighed in Bunsen burner broilers. For this purpose was selected the Crane Bunsen burner broiler with cast iron fingers heated t o a dull red by the flame, thus giving more radiant heat t han other types not so provided though more commonly used. This was done t o provide the most vigorous Bunsen burner competition available as a basis of comparison for the surface combustion burner and the results varied in specific instances, the general average being very close t o a consumption ratio of 2.

This means t h a t surface combustion broiling can be , done with about half the gas now required on the Bun-

sen burner type with feeble radiation and less t han half on others. The quality of the work done is uniformly better though i t must be confessed t h a t opinion differs somewhat here, never going so far, however, as t o charge the surface combustion broiler with inferior work. The intense radiant heat which, of course, is undei zontrol, permits the thing being cooked t o

be placed from 6 inches t o IO inches away from the burner, promptly sears the outside, prevents the escape of flavoring juices and makes the side tha t is cooking not only clearly visible but illuminates i t so well t h a t t he operation can be perfectly carried on in a dark room.

This gas saving of jo per cent or better for radiant operations like the broiling reported above, is con- firmed almost identically by the tests on the radiant room heater, also compared with the T'ulcan which carries a perforated cast iron plate attaining a dull t o bright red by the heat of a Bunsen bar burner below. Heat was here measured by the temperature rise of a measured body of water in a flat sheet metal box placed in front, far enough away t o escape contact with any hot gases. The consumption for equal amounts of heat thus thrown out is almost exactly twice for the Bunsen compared with the surface combustion.

It may be said, therefore, t h a t surface combustion domestic appliances can save in all directly radiant operations about half the gas required by Bunsen appliances and in other cases like top burners and ovens a n average of about 35 per cent in round numbers derived from actual tests in laboratory and kitchen. These figures are for continuous operation and are reduced in some cases by intermittent work such as involves the lighting of a burner for a few minutes and then turning i t out almost as soon as i t gets warm. For such service as this, efficiency is not so important as a quick heat and if less time t h a n one minute be available one present type burner can be retained in each stove for i t , reserving for t he surface combustion burners the large gas consumption work where gas savings are equivalent t o appreciable sums of money.

A t this point t he question naturally arises as t o when t h e saving of gas becomes large enough to pay for the electric current consumed by the fan and in what time the net saving will pay for t he increased cost of surface combustion appliances over Bunsen. The answer t o the last question is t o be found only when i t is known how much the appliance will be used as the fixed charges per hour of use depend on the number of service hours per year; in this respect the case is similar t o plant operation. If a domestic range is t o be used one week per year i t matters little how efficient i t may be and the cheapest thing available in first cost is the thing t o buy. On the other hand it will take a very short time t o pay for the extra cost of an appliance t h a t is used every day and doubly SO

if operated many hours per day, especially if t he service requires considerable gas. Calculations could be given for all sorts of hypothetical cases but as any one can make them for conditions t h a t are important t o him they are here omitted; however, one case is included for illustration.

Suppose a domestic range with a full capacity consumption of IOO cu. f t . per hour were operated three hours per day and three hundred days or goo operating hours per year a t an average rate of jo cu. f t . per hour with ;uct iistribution of various burner service as

INSTALLATIOX O F X E W APPARATUS

8 2 4 T H E J O U R N A L O F I N D U S T R I A L .1ND E N G I N E E R I N G C H E M I S T R Y 1'01. j, NO. IO

corresponds t o an average saving of 40 per cent of what a corresponding Bunsen range would require. I t may also be assumed tha t the latter would cost t he consumer $35 and the former $ j o and t h a t gas costs $1.00 per 1000 cu. f t . and electric current I O

cents per K. W. hour. The surface combustion range would consume per year j o X 3 X 300 = 4j,ooo cu. f t . of gas, while t h e corresponding Bunsen would consume 45,000 + (1.0 - 0.4) = 7 j ,ooo cu. f t . so t h a t t he saving of the former over t he latter is 30,000 cu. f t . , worth $30 per year. From this is to be subtracted the cost of 900 hours of electrical supply, which must be estimated, This can be calculated from the power charts given, and it will be found t h a t for even low efficiencies of fan and motor not over 0. j ampere a t I I O

volts or j j watts should be ample. I n fact actual measurement of the Regina vacuum cleaner set , illustrated on the first range, Figs. 48 and 49, which had a delivery pressure twice and a n air volume about four times what was required, required 92 watts with all burners in operation and 90 with none! proving the excess capacity and the probability of operating with fans and motors designed for the service on less t h a n half an ampere.

Assuming t h a t a t the s tar t no efficient fan and motor be available and t h a t t he absurdly high consumption of 9 0 watts would be required for the necessary I O

cu. f t . of air per minute a t 4-j inches water pressure, the electrical cost of operation would be I O X 901 1000 = 0.1 cent per hour, in round numbers, or $9 .00 per year. This makes the net saving in operation $30.000 - $9.00 = $21.00 per year, and as the excess of first cost was assumed t o be $ ~ j . o o i t would be paid off completely in 600 hours of operation, or if uniformly distributed in 81j2 months.

One short way of disposing of the electrical cost is t o consider t h a t the gas saved by a single top burner GAS SAVING NECESSARY TO JUST P A Y COST O F ELECTRIC CURRENT E X -

PENDED FOR F A N OPERATING SURFACE COMBUSTION APPLIANCE? Cu. f t . gas

per hr. cost- ing same as electric cur-

Electric H.P.' Equivalent Amperes at Current at rent at $1.00 required Equivalents ---- - 10 c. per per 1000

by fan watts 1 I 0 volts 220 volts K. W. hour cu. ft. gas 0.01 7 .46 0.06782 0,03391 0.0746 0 . 7 5 0 .02 14.92 0.13564 0.06782 0.1492 I .49 0 .03 22.38 0.20346 0.10173 0.2238 2 . 2 4 0 . 0 4 29.84 0.27128 0.13564 0.2984 2.98 0 .05 37 .30 0.33910 0.16955 0.3730 3 . 7 3 0.06 44.76 0.40692 0.20346 0.4476 4 . 4 8 0.07 52.22 0.47474 0.23737 0.5222 5 . 2 2 0 . 0 8 59.68 0.54256 0.27128 0.5968 5.9T 0 . 0 9 67.14 0.61038 0.30519 0.6714 6 . 7 1 0 . 1 0 74.60 0.67820 0.33910 0.7460 7 .46 0.11 82.06 . 0,74602 0,37301 0 .8206 8 . 2 1 0 . 1 2 89.52 0.81384 0.40692 0.8952 8.95 0 .13 96.98 0.88166 0.44083 0.9698 9 . 7 0 0 . 1 4 104.44 0,94948 0.47474 1.044 10.44

will more than pay for the electric current t o operate the entire range so t h a t for all burners in operation in excess of one the electrical current costs nothing. Tha t this s ta tement is justified is indicated by the following, assuming the inefficient fan and motor t o be used t h a t costs 0.1 cent per hour t o operate. If the service of a s tandard ~j cu. f t . Bunsen top burner

is preferred by a I O f t . surface combustion burner (ratio 1.j) t he gas saving is j cu. f t . per hour, which is worth j X I O O O ~ I O O = 0 . j cent per hour, a net saving of 0.4 cent per hour. This shows how conserva- tive is the general estimate above, t h a t other gas saving a t one burner will more t h a n defray the expense for current for t he entire range, for under the conditions named the gas saving a t one burner is five times the cost of current for all of them.

I n the preceding table there is given a series of equivalents in gas saving, cu. ft. per hour, the value of which just balances various values of electrical horse power requirements of fan.

C O N C L U S I O S

I t is hoped t h a t thik review of the development of surface combustion will show t h a t it is now possible t o design rather t h a n merely invent apparatus, and t h a t such apparatus as commercial conditions may require may now be produced in no more t ime than is necessary t o decide on the models t o be manufactured and the production of an initial stock. However, there is no intention of leaving the impression t h a t t he work of development is finished for i t is only just fairly started, and should be continued with corresponding improvement in appliances for the next half century.

COLUMBIA UNIVERSITY, N E W Y O R K

METHODS FOR THE EXAMINATION OF NATURAL GAS FOR THE PRODUCTION OF GASOLINE By E. S. MERRIAM AND J. A. BIRCHBY

Received July 18, 1913

The production of natural gas gasoline from the casing head gas of t he oil fields, by compression and cooling, has, within the last five years, grown t o be a very important industry. Through the failure of many plants t o obtain gasoline in paying quantities, i t was early recognized t h a t a preliminary examination of any proposed gas should always be made, and naturally the problem was put up t o the analytical chemist. Any one a t all familiar with the literature on natural gas will realize t h a t positive s ta tements in regard t o the presence and quantities of the higher paraffines occurring in natural gas are not numerous. E thane . propane, and butane were early found in natural gas, bu t no reliable figures could be given as to their amounts. I n fact from combustion da ta alone only the total quantity of paraffin vapors present can be determined, and not which ones.' Generally the results are recorded on the assumption t h a t methane and ethane alone are present.

The exact analysis of a mixture of five hydrocarbon gases does not seem possible, with the means available in an ordinary laboratory. Fractional distillation a t -190' C. and --IZOO C. seems t o have given Lebeau and Damiens2 good results. The use of high pres- sur es in con ne c ti on with the critic a1 t e mp e r a t ur e of the vapors believed to be present might lead t o a more or less correct result. though the solubility of a gas in a liquid hydrocarbon must be taken into account.

1 G. A. Burrell, Bur. of Mines, Bull . 15, 67. 3 Cornpi. r e n d . , 156, 144-7 and 325-7; Chem. Abslracts, 7 . 1338

Design of Surface Combustion Appliances.

Documents

Transcript of Design of Surface Combustion Appliances.