EFFECT OF COWL ANGLE IN THE PERFORMANCE OF …€¦ · Sivakumar et al. (2006) performed numerical...

http://www.iaeme.com/IJMET/index.asp 899 [email protected]

International Journal of Mechanical Engineering and Technology (IJMET) Volume 8, Issue 11, November 2017, pp. 899–909, Article ID: IJMET_08_11_091

Available online at http://www.iaeme.com/IJMET/issues.asp?JType=IJMET&VType=8&IType=11

ISSN Print: 0976-6340 and ISSN Online: 0976-6359

© IAEME Publication Scopus Indexed

EFFECT OF COWL ANGLE IN THE

PERFORMANCE OF SCRAMJET AIR INTAKES

Senthilkumar P, Sharan Chandran M, Sreeja Sadasivan and Govindha rasu N*

School of Mechanical Engineering, VIT, Vellore, India

K Senthil kumar

Mechanical and Industrial Engineering, Higher college of Technology, Muscat,

Muscat-133, Oman

ABSTRACT

A Scramjet engine (Supersonic Combusting Ramjet) entails combustion at

supersonic flow by generating high pressure through recompression (oblique) shock

train. Shock Boundary Layer Interaction (SBLI) is an undesirable phenomenon that

occurs inside the scramjet intake which has a detrimental effect on its performance.

Due to adverse pressure gradient created by SBLI, the flow might be decelerated to

subsonic speeds via normal shock. This leads to unstarting of the engine including

other undesirable phenomena such as high drag and increased localized heating. The

effect of cowl angle is studied on the performance of the scramjet engine in this study.

A scramjet inlet prototype is designed in Solid Works followed by CFD analysis using

ANSYS tool. The effects of cowl angle, static pressure rise, total pressure recovery and

Flow Distortion are brought out.

Keywords: Scramjet intake, SBLI, cowl, ramp and stagnation pressure loss.

Cite this Article: Senthilkumar P, Sharan Chandran M, Sreeja Sadasivan and

Govindha rasu N and K Senthil kumar, Effect of Cowl Angle in the Performance of

Scramjet Air Intakes, International Journal of Mechanical Engineering and

Technology 8(11), 2017, pp. 899–909.

http://www.iaeme.com/IJMET/issues.asp?JType=IJMET&VType=8&IType=11

1. INTRODUCTION

A ramjet engine uses the kinetic energy of air to achieve compression via normal shock wave.

But this pressure rise is associated with a huge stagnation pressure loss and reduction to

subsonic speeds. Also, the high temperature rise creates additional losses due to dissociation

of molecules. A scramjet engine is a variant of a ramjet air breathing jet engine in which

combustion takes place in supersonic airflow. The scramjet is composed of three basic

components: a converging inlet, where incoming air is compressed; a combustor, where

gaseous fuel is burned with atmospheric oxygen to produce heat; and a diverging nozzle,

where the heated air is accelerated to produce thrust. Shock wave- boundary layer interaction

Senthilkumar P, Sharan Chandran M, Sreeja Sadasivan Govindha rasu N and K Senthil kumar


(SBLI) is the problem encountered by the air- breathing propulsion system. The SBLIs cause

boundary layer separation due to the adverse pressure gradient imposed by the shock. These

result in total pressure loss, flow unsteadiness, increased aerodynamic drag and heat fluxes,

fluctuating pressure loads and flow distortion in the intake section leading to the reduction of

the overall propulsive efficiency of the hypersonic vehicle and premature structural fatigue of

aero-structures. The quality of the flow field is also reduced significantly. Flow control

mechanisms have been developed to prevent the boundary layer separation and to keep the

overall propulsive efficiency high even at off-design conditions.

Emami et al. (1995) carried out a parametric study of inlet-isolator performance of a dual

mode Scramjet Engine. To carry out the objective, a two dimensional planar inlet-isolator-

diffuser model was designed and fabricated. The experimental studies were conducted in the

cold flow Mach 4 blowdown facility (M4BDF). The set of inlet-isolator performance data was

obtained from a total of 250 geometric configurations resulting from several interchangeable

and rotating cowls as well as isolator sections of different lengths. Holland et al. (1995)

carried out a computational and experimental study of the internal aerodynamics of a three

dimensional sidewall-compression scramjet intake at Mach 10 and studied the effect of

parameters such as aerodynamic contraction ratio (CR), cowl position and Reynolds number

on its performance. The experiments were performed at in the Langley 31 inch Mach 10

tunnel for three values of contraction ratios (CR=3, 5 and 9), three Reynolds numbers

(Re=0.55x106 per foot, 1.14x10

6 per foot and 2.15x10

6 per foot) and three cowl positions. He

concluded that upon increasing the contraction ratio and decreasing the Reynolds number, the

separation induced by glancing shocks generated by the sidewall leading edges increases. He

also stated that CFD results yielded good agreements with the “on design” conditions while

poor agreements were obtained for highly separated flows. Henckels et al. (2004) carried out

studies on flow phenomenon on scramjet inlets wherein the influence of different inlet

geometries on the heat load was determined. The test was carried out using a hypersonic

blowdown tunnel H2K. Also, influence of sidewalls on the flow field was studied.

Introduction of a 3D sidewall leads to a shock that glances along the ramp surface and

impinges on the boundary layer of the adjacent wall. Also it creates a complex pattern of

shock -shock and shock-boundary layer interactions which are difficult to predict using

commercial available CFD codes. Hence for simplicity, the sidewalls are not included in the

geometry used in this work. Hongjun et al. (2005) provided a method for designing a two

dimensional mixed compression two ramp supersonic inlet to maximize pressure recovery and

matching the inlet mass flow demand. An optimization criterion is defined for on-design

conditions to maximize pressure recovery. For off design conditions, the pressure recovery

depends upon the inlet configuration. A conservative value of total pressure recovery is

calculated to determine the performance of the engine. The mass flow demand is matched by

adjusting the second ramp angle and the bypass door.

Sivakumar et al. (2006) performed numerical simulations for a 3D inlet section using

commercial coding package ANSYS Fluent. The SST k-ω turbulence model was used with y+

value of 105. The results were validated by simulating flow through a 2D mixed compression

hypersonic inlet as well as comparing with experimental data. From the characteristic curves

obtained, the value critical angle of attack was determined to be 8⁰ with higher values of flow

distortion for other values of angle of attack. Krause et al. (2006) conducted an experimental

and numerical analysis on a scramjet inlet using Reynolds Averaged Flow Solvers. Two

intake models were studied, one with sharp leading edge and the other with a blunt leading

edge. Also, various turbulence models were investigated. The Shear Stress Transport (SST)

and Spalart-Allmaras (SA) turbulence models yielded similar results while the original k-ω

turbulence showed much greater separation and thicker boundary layers. The results were not

validated with experimental data and only one result was presented. The result concluded that

Effect of Cowl Angle in the Performance of Scramjet Air Intakes


the side wall shock bends the ramp shock upstream resulting in a separation bubble which

blocks the flow in the isolator. Subash et al. (2011) carried out numerical simulations on the

flow through inlet and isolator of a dual mode scramjet engine at Mach 4 who used the two

equation SST k-ω turbulence model in

ANSYS Fluent. The SST k-ω turbulence model captures boundary layer separation under

adverse pressure gradient effectively and is recommended for problems involving

compressible flow. The results obtained were validated with experimental data for three

different cowl lengths and four different cowl convergence angles. Soumyajit et al. (2012)

carried out numerical simulations using three dimensional Reynolds Averaged Navier Stokes

(RANS) solvers and the SST turbulence model. The steady state simulations were carried out

for different Mach numbers ranging from 3 to 8. Effect of wall boundary conditions – both

isothermal (T=300K) and adiabatic on the starting Mach number was investigated. It was

found that the Mach number for which the inlet unstarts was higher in case of adiabatic wall.

Babinsky et al. (2011) investigated the efficiency of different vortex generators geometries in

presence of a normal shock via oil flow visualisations and determination of separation regions

for different upstream positions. He concluded that micro ramps do not entirely eliminate

boundary layer separation but reduce it to a certain extent. Also, the corner vortices were

found to be enlarged for all cases with boundary layer control – the reason for which is yet to

be determined. Boundary layer separation caused due to SBLI hampers the pressure recovery

in the isolator and in case of high combustor back pressure, leads to unstarting of the inlet. In

his book, “Shock wave Boundary

layer interactions” he has explained the phenomenon of SBLI and its flow physics in great

detail. Saad et al. (2012) carried out experimental to test the effect micro ramps on supressing

the shock boundary layer interactions (SBLI). The freestream Mach number was maintained

at 5 and three different configurations of micro ramps – MR40, MR60 and MR80 were tested.

Detailed flow field behind the micro ramp was investigated using Schlieren imaging and

Infrared thermography. The micro ramp configuration MR80 was found mitigate SBLI better

than the former two. Sun et al. (2012) carried out an experimental study to investigate the

three dimensional flow field behind a single micro ramp in a Mach 2 supersonic flow.

Manoviraj singh et al. (2017) studied the performance of Cesna-172 aircraft. Senthilkumar

and Sreeja (2016) reviewed on high speed flow with micro ramps. Senthilkumar and kamalraj

(2016) studied the numerical simulation of a scramjet intake with micro-cavity.

A pair of streamwise vortices originating from the trailing edge of the ramp enveloped by

an arc shaped external shear layer was captured in the time averaged organization. On

instantaneous analysis, the enveloping vortices were found to be the Kelvin-Helmholtz (K-H)

instability that bounded the wake region. A net lift force is experienced by both type of

vortices.

2. COMPUTATIONAL MODELLING AND MESHING

2.1. Scramjet Geometry

The geometry of the scramjet engine was designed in Solid Works. Initially, the analysis was

done only for a scramjet engine alone without any blockages for cowl angles 0deg, 4deg, 5deg

and 6deg respectively. The computational data obtained for the case of 0 deg was validated

with the experimental values. The data included pressure values at different stream wise

locations and the shock pattern obtained in the inlet-isolator. The dimensions used and the

model are given below in Fig 1 and 2 respectively.



Figure 1 Scramjet Geometry Dimensions

Figure 2 Scramjet Geometry

Figure 3 Dimensions of Computational Domain

In the above Fig 3, the computational domain is shown with the dimensions and

nomenclature of different surfaces. The spanwise width of the geometry is 30mm (origin at

z=0mm) and is axisymmetric about the plane passing through z=15mm. The viewing plane in

the above figure is at z=30mm which was named as ‘side’. For simplification of geometry,

computational time and the resulting three dimensional flow-field, sidewalls are not included

in the analysis. Instead, the surface ‘side’ at z=30mm is considered as a wall with a

symmetrical plane at z=15mm.



2.2. MESHING

A structured grid was used for all the cases simulated in ANSYS Fluent 15.0. The meshing

was carried out in ICEM CFD as shown in Fig 4. To capture the shock boundary layer

interactions, inflation layers and appropriate y+ (the first minimum grid point from the wall)

distance was maintained. For a Reynolds number of 4.8x105, estimated wall distance was

2.9x10-7

to keep the value of y+ less than 1. These values were calculated using y-plus

estimation calculators available. online. The cell number varied from 60,000 for cases without

blockages and micro ramps to 2.1 million for cases with blockages and micro ramps. In each

case, the pre-mesh quality (in Fig 5) was obtained above 0.7 with aspect ratio less than 1000.

Figure 4 Meshing with inflation layers

Figure 5 Pre-mesh quality



2.3. GRID INDEPENDENCY TEST

Grid Independency test (GIT) was carried out on the model (scramjet cowl angle 0 deg) with

both structured and unstructured mesh having same amount of cells – 0.9 million. The details

of each mesh are provided in the table below. The variation in the results was found to be

within ±8%. The meshing images for cases with ramps and blockages including the mesh

report are mentioned in the Appendix. Fig 6 shows the velocity distribution along the mid-

section of the isolator and Fig 7 depicts the percentage error between the velocity values of

hex mesh and test mesh.

Figure 6 Velocity distribution along the Isolator

Figure 7 Percentage error between Hex and Tet mesh

2.4. SETUP IN ANSYS FLUENT

Since the flow is compressible in a supersonic flow, a density based solver is used. The

energy equation is enabled and the k-ω SST turbulence model is used to capture boundary

layer separation accurately under adverse pressure gradient. This model has been widely used

in numerical simulations of supersonic flows owing to the above mentioned advantage. The

operating pressure is kept as 0 Pa (due to presence of vacuum inside the wind tunnel) so that

the absolute pressure equals the gauge pressure. The boundary conditions are as follows:



Table 1 Boundary Conditions used in the setup

The outlet pressure cannot be predicted at the beginning of the calculation as the flow is

supersonic i.e. the conditions at the outlet cannot be communicated upstream (Ref Sec 2.1 and

3.1). Fluent 15.0 calculates the outlet pressure gradually from the inlet conditions as the

calculation proceeds. The turbulence intensity was calculated by the following empirical

relation,

� = 0.16(��)−1∕8 (6.1)

Solution method used is Implicit AUSM solver. Discretization of Second order is used

for increased accuracy. The Courant number was kept at 0.5 and the under relaxation factors

as well as the discretization order were reduced initially. After a stable solution was obtained

(after ~2000 iterations) they were increased gradually. All the residuals were maintained at

1e-04 and to detect convergence, a surface monitor of mass flow rate was enabled at the

outlet. To capture the shock waves effectively, Adaptive grid was used using the pressure-

gradient.

3. RESULTS AND DISCUSSIONS

3.1. Cowl Angle 0 deg

The pressure contours for cowl angle 0 deg is given in Figure 8.

Figure 8 Pressure contours for cowl angle 0 deg at symmetry plane



As seen from Fig 8, a clean shock pattern is observed when the cowl convergence angle is

0 deg. The recompression shock train in the isolator region is also observed distinctly. The

cowl is designed to produce a two-shock system with the second shock impinging just

upstream of the ramp shoulder. The majority of the pressure rise is seen to be occurring from

the first two shocks while remaining fairly constant in the isolator region. Also, high pressure

region at the shock impingement location is observed due to shock-shock interactions

between the incident and reflecting shock which also leads to high localised temperature.

Figure 9 Velocity contours for cowl angle 0 deg at symmetry plane

Figure 10 Shock Induced Boundary Layer Separation at symmetry plane

From figs 9 and 10, shock induced boundary layer separation region is observed on the

ramp surface near the shoulder region with some boundary layer thickening in the isolator

section. Also, due to boundary layer separation, a separation and a reattachment shock is

observed as well. In the isolator region as mentioned earlier, minor boundary layer thickening

due to the pressure gradient of the oblique shock wave train is witnessed. Hence one can

conclude that successive reflections lead to decrease in the shock strength and the pressure

gradient across them.



Figure 11 Pressure distribution at ramp and cowl surface

Figure 11 shows the pressure distribution along the ramp and the cowl surface. On

following the ramp surface (red line) a plateau region is observed which depicts the constant

pressure along the ramp surface. The sudden rise in pressure at X∽90mm is due to the cowl

tip shock. The peaks are observed in the regions of shock-shock interactions. Alternate crests

and troughs evident of the oblique shock train in the isolator. The pressure distribution on the

cowl surface (black line) follows the same profile of the ramp surface.

The pressure contours for cowl angles 4, 5 and 6 deg are found to be similar with structure

of recompression shock train. In these cases, the cowl shock and the ramp shock interact

significantly thereby affecting the pressure rise at throat region. The tabulation of pressure

values at the throat region with other performance parameters will be mentioned shortly in the

report. The velocity contours at the isolator exit are an array of various individual velocity

values. This non-uniformity is caused by the phenomenon call as Flow Distortion. Flow

distortion in a scramjet inlet creates unsteadiness and hampers its pressure recovery. The

variation of flow distortion with respect to the cowl angle is discussed further in this section.

Contours of velocity profile are plotted below to study the effect of ramp and cowl shock

interaction on the flow field of the isolator.

Table 2 Variation of performance parameters with cowl angle

From Table 2 it can be seen that as the cowl angle increases, the pressure recovery

decreases due to increase in the flow distortion. The flow distortion increases due to increase

in the strength of the incident shock due to interaction of the former with the expansion waves

emanating from the ramp shoulder. The inlet is in the ‘start’ condition for all the cowl angles

as Mach number at the isolator end is greater than 1. The plots of the performance parameter

are plotted below.



Figure 12 Static Pressure vs Cowl angle Figure 13 Mach number vs Cowl angle

Figure 14 Pressure Recovery vs Cowl Angle Figure 15 Flow distortion vs Cowl Angle

As depicted in Figs 12 to 15 above, increase in cowl angle decreases the strength of the

incident shock. Separation regions are lesser in the final three cases as the boundary layer is

resistant to the adverse pressure (incipient separation shock). But flow distortion increases

with cowl angle as well which leads to inlet unsteadiness and loss in pressure recovery as

shown above.

5. CONCLUSION

The effect of cowl angle is studied on the performance of the scramjet engine. A scramjet

inlet prototype is designed in Solid Works followed by CFD analysis in ANSYS Fluent 15.0

to study the effects of cowl angle static pressure rise, total pressure recovery and Flow

Distortion. It is noticed that the increase in cowl angle increases the flow distortion. This

leads to inlet unsteadiness and loss in total pressure recovery. On the other hand, the incident

shock strength decreases so that the separation region becomes lesser in size.



REFERENCES

[1] Saied Emami, Carl A. Trexler, Aaron H. Auslender and John P. Weidner, Experimental

Investigation of Inlet-Combustor Isolators for a Dual-Mode Scramjet at a Mach Number

of 4, NASA Technical Paper 3502.

[2] Scott D Holland, Internal Aerodynamics of a Generic Three Dimensional Scramjet Inlet at

Mach10, NASA Technical Paper 3476.

[3] Henckels, A., and P. Gruhn. Study on aerothermal effects of viscous shock interaction in

hypersonic inlets. Fifth European Symposium on Aerothermodynamics for Space

Vehicles. Vol.563. 2005.

[4] Ran, Hongjun, and Dimitri Mavris. Preliminary design of a 2D supersonic inlet to

maximize total pressure recovery. AIAA Paper 2005-7357, Sept 2005

[5] Sivakumar, R., and V. Babu. Numerical simulations of flow in a 3-D supersonic intake at

high Mach numbers. Defence Science Journal 56.4 (2006): 465-476.

[6] Krause, M., B. Reinartz, and J. Ballmann. Numerical computations for designing a

scramjet intake. Department of mechanics, RWTH Aachen University. 25th International

congress of aeronautical sciences. 2006.

[7] Janarthanam, S. and Babu, V. (2012). Numerical simulations of the flow through the inlet

and isolator of a mach 4 dual mode scramjet. The Aeronautical Journal 116(1182), 833–

846.

[8] Saha, Soumyajit, and Debasis Chakraborty. Hypersonic Intake Starting Characteristics–A

CFD Validation Study. Defence Science Journal 62.3 (2012): 147-152.

[9] Manoviraj Singh S, Arjun Singh, Govindha Rasu N, Performance Improvement Study of

CESSNA-172 Aircraft using CFD, International Journal of Mechanical Engineering and

Technology 8(9), 2017, pp. 414–424.

[10] Senthilkumar P, Sreeja Sadasivan, A review on high speed flow with micro ramps,

Journal of Chemical and Pharmaceutical Sciences, Special issue-3, 2016, 39-43.

[11] Senthilkumar P, Kamalraj Thiyagaraja, Numerical simulation of a scramjet intake with

micro-cavity, Journal of Chemical and Pharmaceutical Sciences, Special issue-3, 2016,

35-38.

[12] Rajesh Prabha N, Edwin Raja Dhas J and Ramanan G, Finite Element Structural Analysis

of Connecting Rod of Aa7075-Tic Composite Using ANSYS, International Journal of

Mechanical Engineering and Technology 8(7), 2017, pp. 1102– 1110.

[13] Yousif Khudhair Abbas, Effect of Sideslip Angle on the Balance of Aircraft Moments

Through Steady - State Spin. International Journal of Civil Engineering and Technology,

8(10), 2017, pp. 627–633.

[14] Ali Salah Ameen, Dr. Ajeet Kumar Rai, Analysis of Electronic Chips Microchannel by

Using Ansys Software, International Journal of Advanced Research in Engineering and

Technology (IJARET) ISSN 0976 - 6480 (Print) ISSN 0976 - 6499 (Online) Volume 5,

Issue 7, July (2014), pp. 29-38

EFFECT OF COWL ANGLE IN THE PERFORMANCE OF …€¦ · Sivakumar et al. (2006) performed numerical...

Documents

Transcript of EFFECT OF COWL ANGLE IN THE PERFORMANCE OF …€¦ · Sivakumar et al. (2006) performed numerical...