Auxiliary Inlet Design Study for Mach 1.

4
Layton W. Howerton1 and John W. Slater2
NASA John H. Glenn Research Center, Cleveland, Ohio 44135

Auxiliary inlet configurations were developed for inlets designed for a NASA Mach 1.4 Supersonic
Technology Concept Aeroplane (STCA). The inlets included axisymmetric pitot and axisymmetric
spike inlets. The auxiliary inlets were full circumferential slots within the cowl. Computational fluid
dynamics (CFD) simulations were performed to obtain inlet performance metrics of total pressure
recovery and radial distortion at the engine face for take-off and approach conditions. Methods of
design of experiments were used to explore the statistical significance of the design factors for the
auxiliary inlets which included the axial location, length of the opening, and auxiliary inlet angle. The
results demonstrate the effect of the design factors and show the potential for improved inlet
performance with the use of auxiliary inlets.

Nomenclature
A = Area
D = Diameter
h = Altitude
L = Length
M = Mach number
q = Dynamic Pressure
p = Pressure
T = Temperature
W = Flow rate
α = Angle-of-Attack
ϕ = Auxiliary inlet angle
x, y, z = Cartesian coordinates
( )0 = Freestream property
( )2 = Property at the engine-face station
( )aux = Property of auxiliary inlet
( )SD = Property at the subsonic diffuser station

I. Introduction
The success of commercial supersonic aircraft will depend on the good performance of the various subsystems
of the aircraft. Within the propulsion system, the inlets should provide the required amount of high-quality airflow to
the turbofan engines at all operating conditions of the aircraft. In this paper, we examine inlets for the NASA
Supersonic Technology Concept Aeroplane (STCA) [1] designed as a 55-ton business aircraft capable of cruise at
Mach 1.4. Of special focus will be the aerodynamic performance of the inlets at the take-off and approach conditions
and the effects of auxiliary inlets to provide additional airflow to the engines at this condition.
Section II will discuss the freestream and engine face conditions of the STCA that provide the boundary conditions
for the design of axisymmetric pitot and axisymmetric spike inlets. The inlets were designed using the Supersonic
Inlet Design and Analysis (SUPIN) tool [2]. The auxiliary inlets were modeled with design factors specifying the
location, size of the area opening, and surface angles. The modeling of the inlets and the auxiliary inlets was also
performed within SolidWorks. Section II will also discuss the design and modeling of the inlets with the auxiliary
inlets.

1
Pathways Intern, Propulsion Division / Inlets and Nozzles Branch, AIAA Student Member.
2
Research Aerospace Engineer, Propulsion Division / Inlets and Nozzles Branch, AIAA Associate Fellow.

1
 Section III will discuss the computational methods for the studies. The methods included a quasi-one-dimensional
analysis of the inlet flow rates with the auxiliary inlets. Computational fluid dynamics (CFD) methods are discussed
for the analysis of the flow about and through the inlets. The methods were facilitated using the SolidWorks Flow
Simulation (SWFS) [3] and Wind-US [4] software. The CFD solutions were processed to obtain the inlet performance
metrics as characterized by the inlet flow rates, total pressure recovery, and radial total pressure distortion. To examine
the significance and interactions of the various design factors for the auxiliary inlets, methods of design-of-
experiments (DOE) were used for statistical analysis of the factors and building of surrogate models for the inlet
performance with respect to the design factors.
Section IV will discuss the results of the CFD simulations and DOE analyses for the inlets. The results provide
design information for the inlets of this study, as well as, for inlets being designed for similar commercial supersonic
aircraft.

II. Inlets and Auxiliary Inlets

The inlets were designed for the conditions of the NASA STCA, which is described in the first subsection. The
second subsection discusses the design of axisymmetric pitot and axisymmetric spike inlets using the SUPIN tool.
Configurations for the auxiliary inlet and its associated design factors are discussed in the third subsection.
A. Supersonic Technology Concept Aeroplane
The STCA was developed by NASA to explore several technology goals for commercial supersonic flight [1].
Figure 1 shows an image of the STCA that features three turbofan engines with two engines mounted on the side of
the rear fuselage and one engine mounted within the vertical stabilizer. The image shows axisymmetric spike inlets.

Figure 1. Image of the STCA.

The mission profile for the STCA established the speed and altitude of the aircraft. These provided the freestream
conditions that define the state of the uniform flow approaching the inlets. The inlets were assumed to be isolated
with the freestream designated as station 0. The freestream conditions provided the upstream boundary conditions for
the inlet designs. The freestream was characterized by the Mach number (M0), altitude (h0) or dynamic pressure (q0),
and angle-of-attack (0). The Standard Atmosphere was used to find the static pressure and temperature that defined
the thermodynamic state of the freestream for calculation of the freestream total pressure (pt0) and temperature (Tt0).
Table 1 lists the freestream conditions across the Mach number range. The take-off condition is at M0 = 0.25. The
approach condition is at M0 = 0.30. The cruise condition is at M0 = 1.4.

Table 1. Freestream conditions and turbofan engine-face corrected flow rates and Mach numbers.
M0 h0 (ft) q0 (psf) pt0 (psf) Tt0 (oR) WC2 (lbm/s) M2
0.00 0 0.00 2116.23 518.67 424 0.69976
0.25 0 92.59 2210.27 525.15 420 0.68592
0.30 0 133.32 2252.58 528.01 419 0.68256
0.40 5000 197.21 1966.03 516.87 424 0.69976
0.50 15000 209.00 1416.67 488.44 427 0.71062
0.60 20000 245.07 1240.43 479.56 427 0.71062
0.70 30000 215.55 871.71 452.03 427 0.71062
0.80 35000 223.09 759.06 444.27 427 0.71062
0.90 40000 222.08 662.45 453.14 427 0.71062
1.00 45000 215.61 583.04 467.96 427 0.71062
1.10 48000 225.85 569.34 484.34 427 0.71062
1.20 49000 256.17 616.28 502.28 427 0.71062
1.30 50000 286.54 671.11 521.78 423 0.69623
1.40 50000 332.32 770.79 542.84 413 0.66323

2
 The turbofan engine conceptualized for the STCA was developed based on publicly available data related to the
CFM International CFM56-7B engine. Details of the engine configuration are discussed in Reference [1]. Figure 2
shows images of the engine models for the CFM56 and the STCA turbofan engines. The engine face is designated as
station 2 and modeled as an annular cross-section with an engine-face diameter of D2 = 3.625 feet and a spinner for
the hub with a hub-to-tip ratio of Dhub/D2 = 0.3. The spinner is modeled with an elliptic profile with an aspect ratio of
length-to-diameter of 2.0. The cross-sectional area of the engine face (A2) is the annular area formed by the circular
engine face and spinner hub diameter and is A2 = 9.3918 ft2. The engine flow rate is specified by the engine-face
corrected flow rate (WC2), which is set by the desired level of thrust at a given point in the mission profile. The engine-
face corrected flow rate corresponds to an engine-face mass-averaged Mach number (M2). Both the engine-face
corrected flow rate and Mach number are listed in Table 1 for each of the freestream Mach numbers. The engine-face
geometry and flow rates provide the downstream boundary conditions for the inlet designs.

Figure 2. Images of the models for the CFM56 (left) and STCA (right).
engines.
B. Inlet Designs
The inlets of this study included axisymmetric pitot and axisymmetric spike inlets designed as external-
compression supersonic inlets using the M0 = 1.4 cruise freestream and engine-face conditions as listed in Table 1.
The inlet designs were created using the SUPIN tool [2]. SUPIN is a Fortran 95 program that reads in a text-based
input data file that provides the values of the design factors. SUPIN uses analytic, empirical, and computational
methods to design the inlet and estimate the flow rates, total pressure recovery, and drag for the inlet. SUPIN generates
surface grids for the inlet and writes a Plot3D file [5]. SUPIN can also automatically generate a multi-block, structured
grid for a flow domain about the inlet for flow analysis using CFD methods.
The axisymmetric pitot inlet provided the simplest inlet and served as a baseline for inlet performance. Figure 3
shows images of the axisymmetric pitot inlet. The inlet had an essentially sharp cowl lip modeled with an elliptical
profile with a thickness of 0.002 feet. The cowl interior provided a gradual diffusion from the cowl lip to the engine
face over a distance of one engine-face diameter. SUPIN estimated the total pressure recovery of pt2/pt0 = 0.9557 with
most of the total pressure loss being due to the normal shock with pt1/pt0 = 0.9582. Station 1 was designated as the
cowl lip plane.

Figure 3. Axisymmetric pitot inlet.

The axisymmetric spike inlet introduced a centerbody spike to an axisymmetric cowl. Struts were needed to
support the centerbody. Figure 4 shows images of an axisymmetric spike inlet. The spike consists of a single cone
with a conical half-angle of 13 degrees. The spike angle was also chosen to match the conical spike to a centerbody
with a uniform diameter within the inlet matching the engine face hub. The constant-diameter centerbody was the
result of very little diffusion needed for the subsonic diffuser. The spike resulted in the external supersonic

3
compression to MEX = 1.3 just upstream of the terminal shock located at the cowl lip plane. The conical flow about
the spike results in a flow angle of 1.7 degrees at the cowl lip. The cowl lip interior angle matched this local flow
angle and allowed a cowl lip exterior angle at the start of the cowl exterior of 7.0 degrees. Four support struts were
used for the inlet. The length of the cowl was approximately equal to the engine face diameter, D2. SUPIN estimated
the total pressure recovery of pt2/pt0 = 0.9761.

Figure 4. Axisymmetric spike inlet.

A few options existed for importing the inlet geometry into the computer-aided design software SolidWorks. The
first option suitable for the axisymmetric inlets was to import a list of planar coordinates for the profiles of the cowl
interior, cowl exterior, and spike centerbody. These profiles were converted to curves and then extruded about the
axis-of-symmetry to form the solid models for the centerbody and cowl. For the axisymmetric spike inlet, planar
curves could also be imported to define the support struts. The second option was to import the Plot3D file for the
inlet surface grid into Pointwise and then export an IGES or STL file from Pointwise which could be imported into
SolidWorks. A third option was to export the Plot3D file from Pointwise to SolidWorks and rebuild the model within
SolidWorks.
C. Auxiliary Inlets
The purpose of the auxiliary inlets is to provide additional passages for external airflow into the inlet at take-off
and approach-to-landing conditions for which the engine airflow demand is greater than the freestream capture airflow
rate of the inlet. In this case, the internal inlet pressures are lower than the freestream and the engine is drawing
airflow into the inlet. The modeling approach of the auxiliary inlets for the inlets of this study was to provide an
axisymmetric slot through the cowling of the inlet, as shown in Figure 5. The auxiliary inlets were created in a
SolidWorks sketch. Inside the sketch, parallel lines cutting through the cowl in the correct orientation defined the
sidewalls of the auxiliary inlet. Rounded corners one-quarter the thickness of the cowl were added to the interior and
exterior surfaces of the cowl. The design factors included the axial position of the center of the auxiliary inlet (xaux),
the axial length of the slot (Laux), and the angle of the auxiliary inlet walls (ϕaux).

Figure 5. Auxiliary inlet design parameters (left) and auxiliary inlet for an axisymmetric pitot inlet (right).

4
 III. Computational Methods
The computational methods used to simulate the airflow through the inlets and auxiliary inlets are described in
this section. The methods include an analytical method based on quasi-one-dimensional assumptions, the SolidWorks
Flow Simulation CFD software, and the Wind-US CFD software. The CFD simulations provided for the computation
of the inlet performance metrics are described in this section.
A. Analytic Method
An analytic method was developed to estimate the area of the opening for the auxiliary inlets. It was based on a
quasi-one-dimensional assumption of the continuity of the airflow through the inlet and auxiliary inlets. The method
required specification of the desired Mach number for the inlet at station 1 and the auxiliary inlet opening and estimates
for the total pressure losses through the inlet and auxiliary inlets. A simple solution method yielded the required area
for the auxiliary inlet opening. The method was incorporated into a MATLAB script, as well as into SUPIN and
provided a starting point for the length of the auxiliary inlet for the axisymmetric pitot inlet. The subsequent CFD
simulations indicated that the analytic model underestimated the length and CFD simulations were needed to properly
design the auxiliary inlets.
B. SolidWorks Flow Simulation
SolidWorks Flow Simulation (SWFS) [3] is an add-in program to SolidWorks capable of running CFD simulations
for the models created in SolidWorks. For the CFD simulations, the solid models of the inlets were extended
downstream. Downstream of the engine face, a constant area duct was added to move the internal outflow boundary
condition downstream of the engine face by one engine face diameter. This reduced possible interference of the
internal outflow boundary condition with flow properties at the engine face. A solid disk was created at the engine
face plane, as shown in Figure 6, to monitor the flow properties at the engine face as part of the process for achieving
the desired engine face Mach number for the simulation. The disk was hidden from the flow solver, so there was no
influence of the disk on the flow solution. Downstream of the cowl exterior, a solid cylinder was added to move the
external outflow boundary condition further downstream to reduce interference with the auxiliary inlet flow.

Figure 6. Additional features of the solid model for the SWFS CFD simulations.
The computational flow domain for the simulations of the axisymmetric pitot and axisymmetric spike inlets
consisted of a five-degree wedge at the bottom of the inlet as shown in Figure 7. Due to the subsonic freestream
conditions, the external boundaries were placed four engine face diameters from the cowl lip as to not cause
interference from the boundary conditions on the internal inlet flow.
A Cartesian structured grid was generated for the computational flow domain. The grid defined for a SWFS case
is categorized as global or local. There can only be one global grid, but many local grids can be generated. Local grids
can be placed in a volume of space or be defined by the shape of a solid body in the model. The global grid defines
the number of grid points to be placed in the finite volume of the computational domain. The global grid can also
include an aspect ratio factor to create the largest cells away from the model and smallest cells nearest to the model.
The local grid was used in this study to resolve the boundary layer flow near the inlets. Both local and global grids are
controlled with sliders to determine the characteristics of the grid. Grid generation of the Cartesian grid is then
completed for both local and global based on the combination of slider positions. Options are available in SWFS for
automatic grid refinement while running SWFS cases, but this option was not used in this study.

5
 SWFS solves the Favre-averaged Navier-Stokes equations. A perfect gas model was used for the air in the
computational domain. Turbulence was modeled in the simulations using the k-ε turbulence model. Additionally, a
wall function was used to resolve thick and thin boundary layers. A grid study was performed to refine the grid near
the inlet surfaces to provide sufficient resolution for the wall functions. The freestream conditions were imposed on
the inflow and external outflow boundaries. An internal outflow boundary condition was set to specify the outflow
Mach number for the internal flow. The Mach number, total pressure, static pressure, and mass flow rate were
monitored by SWFS at the engine face disk to determine iterative convergence of the simulations. Iterative
convergence was considered achieved once stable values for all the engine face variables were achieved for one
hundred iterations.

Figure 7. Computational flow domain for the axisymmetric inlets.

C. Wind-US CFD Solver

The Wind-US flow solver [4] solves the steady-state, Reynolds-averaged Navier-Stokes (RANS) equations for the
flow properties at the grid points of a multi-block, structured grid defining a flow domain about the inlets. Wind-US
used a cell-vertex, finite-volume representation for which the flow solution was located at the grid points and a finite-
volume cell was formulated about the grid point. In Wind-US, the RANS equations were solved for the steady-state
flow solution using an implicit time-marching algorithm with a first-order, implicit Euler method using local time-
stepping from an initial flow solution. All the simulations were performed assuming calorically perfect air. The
inviscid fluxes of the RANS equations were modeled using a second-order, upwind Roe flux-difference splitting
method. The flow simulations assumed fully turbulent flow in which the turbulent eddy viscosity was calculated using
the two-equation Menter Shear-Stress Transport (SST) [6] turbulence model.
Figure 8 shows the computational flow domain and boundary conditions (BC) used for the planar CFD simulations
of the axisymmetric pitot inlets. The flow domain defined the control volume in which the RANS equations were
solved. The internal and external surfaces of the inlet formed a portion of the boundary of the flow domain where
non-slip, adiabatic viscous wall boundary conditions were imposed. The inflow, far-field, and external outflow
boundaries of the flow domain had freestream boundary conditions imposed in which the Mach number, pressure,
temperature, and angle-of-attack were specified. The inflow, far-field, and external outflow boundaries were
positioned about five engine-face diameters from the cowl lip.
Downstream of the engine face, a constant-area outflow nozzle section was added to the flow domain to set the
flow rate through the inlet. The nozzle section moved the internal outflow boundary condition downstream of the
engine face, which reduced possible interference from the boundary condition on the flow at the engine face. The
outflow nozzle is shown in Figure 8. The length of the outflow nozzle section was about 1.7 times the diameter of the
engine face. At the internal outflow boundary, the static pressure, Mach number, or corrected flow rate was specified
as the boundary condition.

6
 Figure 8. Flow domain and boundary conditions for the axisymmetric pitot inlet CFD simulations.

The computational grid for the flow domain was generated by dividing the flow domain into multiple blocks and
generating structured grids for each block. SUPIN was used to generate the blocks and grid points using an automated
process. SUPIN also created the boundary condition file for Wind-US. The inputs to the process include some factors
to determine the extent of the flow domain and the resolution of the grid points. The grid resolution factors include
the grid resolution of the first grid point away from the wall (swall) and the grid resolution within the throat section
in the streamwise direction (sx). The wall spacing (swall) was set to resolve the boundary layer with y + = 1. SUPIN
then imposed these grid resolution values along the edges of the inlet geometry and flow domain to compute the
required number of grid points along those edges. A grid block topology was assumed for the inlet to form the edges
into faces and those faces into blocks. SUPIN generated grids along the edges, on the surfaces, and within the interior
volume of each block. The interior block boundaries abutted with other block boundaries. For most blocks, the grid
lines were contiguous across block boundaries, but some non-contiguous boundaries were used to facilitate the
structured topology. Figure 9 shows examples of the grid lines for the planar grid.
The CFD simulations were initialized with a flowfield set to the freestream conditions. The simulations were
started with the first-order form of the Roe flux-splitting method to damp out large initial gradients. Eventually, the
second-order flux method was applied as the residuals over the iterations decreased and the boundary layers, shock
waves, and subsonic inlet flow took form. At the start of the simulations, the Courant-Friedrichs-Lewy (CFL) number
had a value of 0.5 but increased incrementally to a value of 2.5 as the flow solution developed. Local time stepping
was used in which the local time step used was computed based on the CFL number and the local grid cell size. The
iterative convergence was indicated in part by the reduction of the root-mean-square of the residuals of the
conservative variables for each block. Iterative convergence was also evaluated through the monitoring of the
convergence of the inlet flow rate, total pressure recovery, and total pressure distortion. The steady-state solution was
considered converged when these values varied less than 0.1% of their values over 1000-2000 iterations.

7
 Figure 9. Structured, multi-block, computational grids on the symmetry plane for
the axisymmetric pitot inlet.

D. Inlet Performance Metrics

The flow solutions from the CFD simulations were used to obtain the aerodynamic performance metrics for the
inlet. The three inlet performance metrics used to characterize the inlet included the inlet flow ratio (W2/Wcap), the
inlet total pressure recovery (pt2/pt0), and the SAE ARP 1420 radial (DPRP) distortion index. The inlet flow ratio was
defined as the inlet flow rate (W2) divided by the reference capture flow rate (Wcap). The inlet flow rate (W2) was
obtained from the CFD simulation by integrating the rate of flow passing through the cross-stream grid planes of the
outflow nozzle. The total pressure at the engine face (pt2) was computed as the mass-average of the total pressures at
the grid plane at the engine face. The inlet radial (DPRP) total pressure distortion index at the engine face was defined
on a standard 40-probe rake array of the Society of Automotive Engineers (SAE) Aerospace Recommended Practices
(ARP) 1420 document [7]. For the planar CFD domain, the rake array consisted of a single rake arm containing five
total pressure probes. Each probe was placed radially at the centroid of equal-area sectors of the annular engine face.
The flowfield from the CFD simulation was interpolated onto the locations of the probes to obtain the total pressure
at the probe location.
E. Design of Experiments
The methods of design-of-experiments (DOE) were applied to statistically examine the three factors for the
auxiliary inlet inlets that influenced the inlet performance. The three factors were the position (xaux), the angle (ϕaux),
and the length (Laux) of the auxiliary inlet. The responses were the inlet performance metrics of inlet flow ratio, inlet
total pressure recovery, and SAE ARP radial distortion index. The primary DOE design used for this work was the
Central-Composite Face-Centered (CCF) design. Three levels of the three factors were used which required 15-factor
combinations to build a quadratic model. Each combination involved one CFD simulation at the specified corrected
flow rate for the M0 = 0.25 take-off condition. To obtain the simulation at the desired inlet flow rate, several
simulations were required while iterating the internal outflow boundary condition. The Design Expert [8] software
was used to perform the statistical analysis.

8
 IV. Results
This section discusses the results of CFD simulations that explored the design factors for the auxiliary inlets. The
results from the DOE studies will be discussed separately for each inlet.
A. Axisymmetric Pitot Inlet
The axisymmetric pitot inlet provided a baseline and simpler geometry for examining the auxiliary inlet
performance. Figure 10 shows Mach number contours and the inlet performance with respect to the range of
freestream Mach numbers. At M0 = 1.4, the cruise design condition, the normal terminal shock stood at the cowl lip
plane. Most of the total pressure loss was due to the normal shock. At M0 = 1.0, the terminal shock was not present,
and the total pressure recovery was nearly unity. At M0 = 0.6, there were some total pressure losses from boundary
layer separation at the sharp cowl lip interior as air was being drawn into the inlet. At M0 = 0.25, there was significant
separation at the sharp cowl lip and portions of the core inlet flow reached supersonic conditions.

Figure 10. Mach number contours on the symmetry plane and inlet performance
metrics of the axisymmetric pitot inlet at various freestream Mach numbers.
An auxiliary inlet consisting of a circumferential slot just ahead of the engine face was included in the
axisymmetric pitot inlet. The overall effect of the auxiliary inlet can be seen in the Mach number contours shown in
Figure 11. The freestream and engine face conditions corresponded to the M0 = 0.25 conditions. With no auxiliary
inlet, the airflow being drawn into the inlet separated at the sharp leading edge and created a separation bubble
downstream of the cowl lip on the interior of the cowl. The auxiliary inlet provided an additional passage for airflow
and the Mach number of the flow entering the cowl lip station was reduced. The engine face total pressure recovery
for the simulation of the inlet with the auxiliary inlet was pt2/pt0 = 0.952 compared to pt2/pt0 = 0.902 without. The radial
distortion decreases slightly with the use of the auxiliary inlet.

Figure 11. Mach number contours for planar CFD simulations of the
axisymmetric pitot inlet without (top) and with (bottom) an auxiliary inlet.
A DOE study was performed to examine the variation of the inlet performance metrics with respect to the three
auxiliary inlet design factors. Fifteen combinations of the input factors were used for the study and the resulting
performance metrics were obtained from CFD simulations performed using SWFS. The auxiliary inlet location was

9
varied from 1.665 to 2.319 feet from the cowl lip. The auxiliary inlet angle was varied from 0 to 40 degrees, and the
length of the auxiliary inlet was varied from 0.127 to 0.211 feet. The resulting total pressure recovery varied from
0.940 to 0.960. The radial distortion varied from 0.118 to 0.347. Design Expert was used to perform the statistical
analysis of the data. With regards to the total pressure recovery, a linear model was suggested to fit the data. The first-
order term for each of the three design factors was included. No terms dictating an interaction of design factors were
kept in the model. The location of the auxiliary inlet, xaux, had a smaller impact on increasing total pressure recovery
than the angle, ϕaux, or length of auxiliary inlet, Laux. The angle, ϕaux, was the dominant design factor. The statistical
analysis showed that increasing any of the design parameters would increase the total pressure recovery. Increasing
the factors translates to a larger angle on a longer auxiliary inlet closer to the engine face plane. The argument agrees
with the logic of increasing flow in the slowest region to improve recovery.
In terms of radial distortion, a linear model was used to fit the data. The length, Laux, of the auxiliary inlet was
deemed statistically insignificant when compared to the other two parameters. No mixed terms for the interaction of
design variables were significant. The angle, ϕaux, was determined to be the dominant factor for radial distortion. To
decrease the radial distortion, the angle of the inlet should be increased, and the location should be moved further
forward.
Optimization of both radial distortion and total pressure recovery showed that the largest angle which was furthest
aft on the longest inlet was the most desirable. This configuration corresponded to the maximum of each design factor
examined for the axisymmetric pitot DOE. This suggested that the optimum was located outside of the bounds
examined. Further examination could verify this prediction.
A CFD simulation was performed of the optimum auxiliary inlet configuration as determined by Design Expert as
a means to verify the optimum configuration. The predicted total pressure recovery was 0.9567, and the CFD solution
gave a total pressure recovery of 0.9600. The radial distortion prediction was 0.1946 with the CFD giving a distortion
of 0.1182.
An additional DOE study was performed using Wind-US on the axisymmetric pitot inlet. The Wind-US DOE
replicated the geometries from the DOE performed in SWFS. A modified aft cowl lip on the auxiliary inlet was
developed to direct the flow more smoothly. An elliptical profile with an aspect ratio of approximately four was also
modeled. Design Expert was used to analyze the solutions. For this additional DOE, the total pressure recovery varied
from 0.918 to 0.935, and the radial distortion varied from 0.2105 to 0.2787.
Total pressure recovery was modeled in design expert as dependent on all three design factors as well as a second-
order position term and mixed terms of position times length and of length times angle. The mixed terms indicate an
interaction or coupling of responses of multiple design factors. As with the SWFS DOE, an increase in any design
factor would lead to an increased total pressure recovery.
Radial distortion in the inlet was modeled to depend on first and second-order terms related to positioning the inlet,
a first-order term related to the length of the auxiliary inlet opening, and a mixed term including first-order terms for
length multiplied by position. The angle was determined to not have a significant influence. This contrasts with the
SWFS DOE conclusion for the axisymmetric pitot inlet that the angle of the auxiliary inlet dominated. The inclusion
in the Wind-US DOE of the elliptical profile on the aft portion of the auxiliary inlet could have eliminated the
effectiveness of the change in angle on radial distortion.
The statistical optimum of the three design factors was determined in Design Expert, and this configuration was
simulated in Wind-US after the completion of the DOE study. The Design Expert optimum had a position of 1.665
feet, an angle of 40 degrees, and a length of 0.211 feet. Agreement was found of a half percent between the model
predicted total pressure recovery and the CFD total pressure recovery for the optimum configuration. The DOE
optimum also showed good agreement between model optimum and CFD in radial distortion.
Comparison between the Wind-US and SWFS DOE studies for the axisymmetric pitot inlet show several key
similarities and differences. The first difference found between the two DOE studies was a discrepancy between nearly
all cases in terms of the specific values for total pressure recovery and radial distortion. The values found by SWFS
for all cases was on the order of 3% higher total pressure recovery, and a margin lower radial distortion. The margin
by which the radial distortion varied from geometry to geometry. However, the trends remained consistent, especially
for total pressure recovery. The qualitative predictions align well with the total pressure recovery for this reason. The
recommendations mentioned in both DOE studies were very similar. The recommendations for the radial distortion
as well as the qualitative analysis regarding the importance of auxiliary inlet angle could be clouded slightly by the
elliptical aft lip included in the Wind-US DOE.
B. Axisymmetric Spike Inlet
The axisymmetric spike inlet provided a higher-performing inlet for a Mach 1.4 aircraft. Figure 12 shows Mach
number contours and the inlet performance with respect to the range of freestream Mach numbers. At M0 = 1.4, the

10
cruise design condition, the normal terminal shock stood at the cowl lip plane. Most of the total pressure loss was due
to the normal shock; however, the pre-shock Mach number was lower than for the axisymmetric pitot inlet, and so,
the total pressure losses were less for the axisymmetric spike inlet. At M0 = 1.0, the terminal shock was not present,
but there was a small region of supersonic flow at the shoulder of the centerbody. At M0 = 0.6, the inlet flow was
choked as air was being drawn into the inlet. At M0 = 0.25, there was significant separation at the sharp cowl lip and
the flow in the inlet was choked. A normal shock formed within the inlet, leading to greater total pressure losses and
distortion.

Figure 12. Mach number contours on the symmetry plane and inlet performance
metrics of the axisymmetric spike inlet at various freestream Mach numbers.
An auxiliary inlet consisting of a circumferential slot just ahead of the engine face was included in the
axisymmetric spike inlet. The first set of CFD simulations considered a planar flow domain that did not include the
struts. The effect of the auxiliary inlet can be seen in the Mach number contours shown in Fig. 13. The freestream
and engine face conditions correspond to the M0 = 0.25 conditions. The auxiliary inlet provided an additional passage
for airflow and the Mach number of the flow entering the cowl lip station was reduced. Compared to the corresponding
frame of Fig. 12, the normal shock was eliminated, and the flow separation at the engine face was reduced. The engine
face total pressure recovery for the axisymmetric spike inlet with the auxiliary inlet was pt2/pt0 = 0.9329 compared to
pt2/pt0 = 0.888 without the auxiliary inlet. Radial distortion was reduced with the use of an auxiliary inlet from 0.235
to 0.172.

Figure 13. Mach number contours on the symmetry plane of the axisymmetric spike inlet
with an auxiliary inlet at a take-off conditions with a freestream Mach number of M0 = 0.25.
Preliminary 3-Dimensional CFD simulations have been performed in SWFS to include the axisymmetric spike
strut. These simulations include a domain that includes a full strut and half of an auxiliary inlet. The auxiliary inlet
ends before reaching the support strut, so the focus of the simulation will be the interaction between the flow entering
the auxiliary inlet and the support strut.
A fifteen case DOE study was performed for three design factors in a similar fashion as the axisymmetric pitot
DOE. The auxiliary inlet location was varied from 1.822 to 2.728 feet from the cowl lip. The auxiliary inlet angle was
varied from 0 to 40 degrees, and the length of the auxiliary inlet was varied from 0.130 to 0.217 feet. The resulting
total pressure recoveries varied from 0.941 to 0.958. The radial distortion varied from 0.193 to 0.335. Design Expert
was used to perform the statistical analyses. A quadratic model was developed for the total pressure recovery response
to the DOE design factors. All first-order terms for each variable, the second-order term for xaux, as well as a mixed
angle and location term were all included in the quadratic equation for the recovery. The second- order term of position

11
dominated the model behavior. The mixed term produced the smallest influence. As with the axisymmetric pitot inlet
statistical model, the model predicts that total pressure recovery will increase as each of the factors is increased. The
mixed term is the only exception as the model predicts that if the mixed term decreases the total pressure will increase.
For radial distortion, a similar model to the axisymmetric pitot was produced. A linear model including first-order
terms for position and angle but not length was created. The length was deemed to have insignificant influence on
distortion in the axisymmetric spike inlet. Position and angle had similar control over the model behavior. As with the
axisymmetric pitot, to decrease radial distortion the angle should be large, and the auxiliary inlet position should be
further forward from the engine face plane.
For the optimization of total pressure recovery and radial distortion, a configuration of maximized angle and length
but minimized position was produced. The model predicted optimum and corresponding case from the axisymmetric
spike DOE represent the same geometry. The predicted total pressure recovery was 0.9542 and the CFD simulation
total pressure recovery was 0.9540. This served as a check that the model predicted optimum was very close to the
CFD solution. The radial distortion prediction was 0.1805 with the CFD giving a distortion of 0.1953.
The optimum CFD solution compared to the solution of a no auxiliary inlet case again shows benefits to including
an auxiliary inlet. Without an auxiliary inlet, the recovery for the axisymmetric spike inlet was 0.9358. The auxiliary
inlet improved the recovery to 0.9540. The radial distortion was decreased from 0.2707 to 0.1953 with the auxiliary
inlet providing additional flow at the cowl.
An additional DOE study was performed using Wind-US. This study repeated the 15 simulations performed in
Solidworks Flow Simulation. A modified aft cowl lip on the auxiliary inlet was used to reduce radial distortion. The
elliptical auxiliary inlet aft lip can be seen in Figure 14 below. Some simulations within the DOE displayed unsteady
behavior. Five simulations were reduced from second-order to first-order solutions to obtain a steady result. Design
Expert was used to analyze the solutions. The results from the additional DOE varied in total pressure recovery from
0.922 to 0.940. The radial distortion varied from 0.250 to 0.365.
Total pressure recovery was determined to be dependent entirely on the length of the auxiliary inlet. The statistical
analysis determined that the influence of the angle of the auxiliary inlet, as well as the location, were not significant.
The trend displayed that a larger auxiliary inlet would improve the recovery performance of the inlet.
Radial distortion in the inlet was modeled to depend on first and second-order terms related to positioning the inlet,
and a first-order term related to the length of the auxiliary inlet opening. The angle was determined to not have a
significant influence. Qualitatively, the inlet distortion is reduced if the additional flow from the auxiliary inlet is
introduced further from the engine face plane.
The statistical optimum of the three design factors was determined in Design Expert, and this configuration was
simulated in Wind-US. The Design Expert optimum had a position of 1.978 feet, an angle of 20 degrees, and a length
of 0.217. There was good agreement between the model predicted total pressure recovery and the CFD total pressure
recovery for the optimum configuration. The predicted total pressure recovery was 0.9391, and the CFD simulation
provided a total pressure recovery of 0.9329. The predicted radial distortion was higher than the CFD simulation
result; a predicted value of 0.2704 compared to 0.1720.
Comparison between the Wind-US and SWFS DOE studies show several key similarities and differences. The
differences between models, similarities in trends, and potential causes will be laid out. When analyzing the DOE
models, the angle was deemed insignificant in the Wind-US DOE but significant by the SWFS DOE. One potential
explanation for this change could be the modified aft lip of the auxiliary inlet used in the Wind-US study. The main
difference between studies is Wind-US predicted lower total pressure recovery and higher distortion than SWFS for
all cases. The cases from Wind-US predicted a total pressure recovery an average of 0.0159 lower, and a radial
distortion 0.0473 higher on average. While the target values were shifted, both DOE studies developed similar
recommendations to increase recovery and decrease distortion. The two studies agreed that increasing the length of
the auxiliary inlet would increase total pressure recovery. Additionally, the dominating factor for radial distortion in
both studies was the axial location. Moving the axial location away from the engine face plane reduced the radial
distortion.

12
 Figure 14. Mach number contours on the symmetry plane of the axisymmetric
spike inlet with an auxiliary inlet at a freestream Mach number, M0= 0.25.
Upon the conclusion of the four DOE studies, an additional investigation was done into redesigning the thickness
of the cowl and a translated cowl on the axisymmetric spike inlet. The DOE simulations suggested that performance
would increase as the auxiliary inlet was moved forward and increased in length. Several simulations were performed
with the maximum cowl thickness varied from about 1 inch to about 3 inches. The additional thickness of the cowl
would allow greater room for the inlet mechanical hardware and acoustic lining within the inlet. The translation of the
cowl forward as opposed to making an opening within the set length of the inlet allowed for greater auxiliary inlet
lengths. The longer auxiliary inlet promised improved aerodynamics, and the translation forward allowed for the
positioning of the auxiliary inlet to be moved forward. The forward position also allows for more area for acoustic
liners. Simulations were completed at take-off, cruise, and approach on the thicker cowl model.
SUPIN design factors were varied and Wind-US simulations were completed to arrive at the three-inch cowl
thickness and one-foot auxiliary inlet length for take-off conditions. Wind-US calculations were then completed with
an auxiliary inlet at take-off, and with and without an auxiliary inlet at cruise and approach to landing. Because of
limitations of SUPIN and Wind-US, the take-off with auxiliary inlet, and approach with auxiliary inlet simulations
were repeated with a translated forward cowl in SWFS. An image of the translated geometry is shown below in Figure
15 alongside an image of the untranslated SUPIN model. The constant-area downstream duct section shown in the
translated model between the blue lines is not a part of the inlet geometry. In Figures 16 and 17, the Mach contour
results are shown for the take-off condition. The much larger 1-foot slot produces much better radial distortion and
total pressure recovery than previously attained. Wind-US provided a total pressure recovery of 0.989 for the
simulation shown in Figure 16. In contrast, during the DOE study, the optimum was found to be 0.932. The same
simulation saw a drop in radial distortion from 0.172 to 0.026. The larger auxiliary inlet produced significant
aerodynamic improvements.

Figure 15. Cowl thickness redesign shown translated forward (right) and
untranslated (left) with 1-foot auxiliary inlet for take-off conditions.

13
 Figure 16. Wind-US untranslated cowl redesign Mach contours with
1-foot auxiliary inlet for take-off conditions.

Figure 17. SWFS translated cowl redesign Mach contours with

1-foot auxiliary inlet for take-off conditions.
For approach to landing, additional SUPIN design and Wind-US simulations were completed to choose a 0.3 foot
length auxiliary inlet to provide the appropriate amount of flow for the reduced engine throttle settings during approach
to landing. The translated and untranslated models for the approach to landing can be seen in Figure 18. The constant-
area downstream duct section shown in the translated model between the blue lines is not a part of the inlet geometry.
In Figures 19 and 20, the Mach contour results are shown for the approach condition. At the reduced engine face Mach
number required at approach to landing, the inlet requires much less auxiliary flow. Thus, the slot length is 0.3 feet.
Approach condition were not simulated in the DOE studies, so there is not a point of comparison except between
Wind-US and SWFS. Wind-US provided a total pressure recovery of 0.993 for the simulation shown in Figure 19.
The SWFS also gave a result of 0.993 for total pressure recovery. Wind-US had a radial distortion of 0.0218 and for
SWFS a value of 0.0111. The agreement of both simulations verifies this configuration should provide good
performance at approach to landing.

14
 Figure 18. Cowl thickness redesign shown translated forward (right) and untranslated
(left) with 0.3-foot auxiliary inlet for approach-to-landing conditions.

Figure 19. Wind-US untranslated cowl redesign Mach contours with 0.3-foot auxiliary
inlet for approach-to-landing conditions.

Figure 20. SWFS translated cowl redesign Mach contours with 0.3-foot auxiliary inlet for
approach-to-landing conditions.

15
 V. Conclusion
Inlets designed in SUPIN were modeled in SolidWorks. SolidWorks Flow Simulation and Wind-US were used to
analyze the flow through auxiliary inlets in the SUPIN models. Statistical analysis in Design Expert was completed
on the resulting data. Consistent recommendations regarding how to best reduce radial distortion and increase total
pressure recovery were developed. Statistically significant improvements to inlet performance with the inclusion of
the auxiliary inlet align with the expected result. The cowl thickness redesign, as well as translation option, represent
the culmination of the lessons learned throughout the design process. A few of the most important qualitative lessons
will be shared here. Moving auxiliary inlets forward produces a better aerodynamic performance as the flow has time
to settle and become uniform at the engine face plane. The thicker cowls that allow for the possibility of mechanical
translation and acoustic lining also help limit flow separation. The larger auxiliary inlets help provide a more uniform
auxiliary flow and increase performance. The auxiliary inlets discussed provided improved flow quality at the engine
face of the SUPIN inlets designed for the STCA engines.

Acknowledgments
This work is supported by the NASA Commercial Supersonic Technology Project.

References
[1] Berton, J. J., Huff, D. L., Seidel, J. A., and Geiselhart, K. A., “Supersonic Technology Concept Aeroplanes for
Environmental Studies,” AIAA Paper 2020-0263, January 2020.
[2] Slater, J. W., “SUPIN: A Computational Tool for Supersonic Inlet Design,” AIAA Paper 2016-0532, January
2016.
[3] Dassault Systemes, “Technical Reference SolidWorks Flow Simulation 2019”, 2019.
[4] Yoder, D. A., ‘Wind-US User’s Guide: Version 3.0,” NASA TM 2016-219078, March 2016.
[5] Walatka, P. P., Buning, P. G., Pierce, L., and Elson, P. A., “PLOT3D User’s Manual,” NASA-TM-101067, March
1990.
[6] Menter, F. R., “Two-Equation Eddy-Viscosity Turbulence Models for Engineering Applications,” AIAA Journal,
Vol. 32, No. 8, 1994, pp. 1598–1605. doi: 10.2514/3.12149.
[7] Society of Automotive Engineers (SAE), “Gas Turbine Engine Inlet Flow Distortion Guidelines,” SAE ARP
1420, Rev. B, February 2002.
[8] Design Expert, “Version 7.0.0. Stat-Ease,” Design Expert Inc., Minneapolis, 2005.

16