Article 1

The experiments and stability analysis of hypersonic boundary 2

layer transition on a flat plate 3

Yanxin Yin 1,2, Yinglei Jiang 3,4, Shicheng Liu3,4 and Hao Dong 3,4,* 4

1 Department of Mechanics, School of Mechanical Engineering, Tianjin University, Tianjin 300072, China. 5
2 China Academy of Launch Vehicle Technology, Beijing 100076, China. 6
3 College of Aerospace Engineering, Nanjing University of Aeronautics and Astronautics, Nanjing 210016, 7
China. 8
4 Key Laboratory of Unsteady Aerodynamics and Flow Control, Ministry of Industry and Information Tech- 9
nology, Nanjing University of Aeronautics and Astronautics, Yudao Street 29, Nanjing, Jiangsu 210016, 10
China. 11
12
* Correspondence: donghao@nuaa.edu.cn 13

Abstract: Experimental and Linear Stability Theory (LST) investigation of boundary layer transition 14
on a flat plate was conducted with a flow of Mach number 5. The temperature distributions and 15
second mode disturbances on the flat plate surface at differen unit Reynolds number (Reunit) are 16
captured by infrared thermography and PCB technology, respectively, which revealed the transi- 17
tion location of the flat plate boundary layer. The results indicate that Reunit has an obvious influence 18
on transition Reynolds number (ReT) in the hypersonic boundary layer. The evolution characteristics 19
of the second mode instability in the flat plate boundary layer are also investigated by LST. With 20
the increase of the Reunit, the thickness of boundary layer decreased, and the growth rate of Mark 21
mode become larger gradually. The amplitude amplification factor (N-factor) of the second mode 22
instability is evaluated by eN method. The N-factor of the transition location in the wind tunnel ex- 23
periment predicted by LST is about 0.98 and 1.25 for Reunit =6.38×106 and 8.20×106, respectively. 24

Keywords: Hypersonic boundary layer; transition; LST; eN method; infrared thermography 25

26

1. Introduction 27
The flow regime within the boundary layer, whether laminar or turbulent, signifi- 28
cantly influences surface friction and heat exchange. Turbulent boundary layers exhibit 29
Citation: To be added by editorial
friction and heat flux approximately 3~5 times higher than those observed in laminar flow 30
staff during production.
[1]. Accurate and efficient prediction of boundary layer transition is crucial for enhancing 31
Academic Editor: Firstname Last- the safety and payload capacity of hypersonic vehicles. Consequently, the prediction of 32
name hypersonic boundary layer transition plays a pivotal role in the meticulous design of the 33
Received: date entire vehicle, particularly in the context of the thermal protection system (TPS). How- 34
Revised: date ever, owing to the inherent highly nonlinear nature, hypersonic boundary layer transition 35
Accepted: date remains a pressing challenge in fluid mechanics [2][3]. 36
Published: date Different from incompressible boundary layer, the two mechanism that affects the 37
hypersonic boundary layer transition are the second mode instability and crossflow in- 38
stability. The second mode instability was first discovered by Mack [4] [5] using the linear 39
Copyright: © 2023 by the authors. stability analysis method to study the hypersonic flat plate boundary layer. Then the sec- 40
Submitted for possible open access ond mode has been measured in subsequent experiments and the results agree well with 41
publication under the terms and theoretical LST predictions [6~8]. At present, wind tunnel experiment still plays an im- 42
conditions of the Creative Commons portant role in the hypersonic boundary layer transition studies. Zhao [9] et al. studied 43
Attribution (CC BY) license the influence of different total temperatures on the basic flow and transition position of 44
(https://creativecommons.org/license the boundary layer under the same Mach number and unit Reynolds number inflow 45
s/by/4.0/).

Appl. Sci. 2023, 13, x. https://doi.org/10.3390/xxxxx www.mdpi.com/journal/applsci

Appl. Sci. 2023, 13, x FOR PEER REVIEW 2 of 13

conditions and pointed out the important significance of ensuring consistent total temper- 46
ature of the inflow in wind tunnel experiments. With the progress of measurement tech- 47
nology, some instantaneous measurement methods, which have high spatial and tem- 48
poral resolution, such as PCB [10~11] and ALTP [12], are used for identifying disturbance 49
modal in the boundary layer. Infrared thermography appears to be a potent tool for in- 50
vestigating boundary layer transitions. Borg et al. [13] employed IR technology to study 51
the boundary layer transition of HIFiRE-5b at a free-stream Mach number of 6. The find- 52
ings suggest that IR can clearly depict high and low-temperature streaks in the streamwise 53
direction on the model surface and capture the transition process. In addition to surface 54
disturbance measurements, some flow visualization techniques are used to capture the 55
process of boundary layer transition. Zhao et al. [14] successfully captured the process of 56
the second mode wave growth and breaking into turbulence in the boundary layer of a 57
conical surface using NPLS technology. Zhu et al. [15] used Rayleigh scattering technique 58
successfully captured second modes of rope waves in the boundary layer of flared cone, 59
and the amplitude of second mode decays before transition to turbulent. Dong et al. [16] 60
accurately measured the wall friction resistance of hypersonic plate boundary layer by 61
using oil film interferometry and determined the transition position according to the wall 62
friction. In addition to wind tunnel experiments, due to the development of computer 63
computing power, many researchers began to use direct numerical simulation (DNS) to 64
study the sensitivity of hypersonic boundary layer of flat plate and the propagation of 65
disturbance [17~18]. 66
Various theoretical analysis techniques, including Linear Stability Theory (LST) and 67
Parabolized Stability Equations (PSE), have been employed for the investigation and pre- 68
diction of hypersonic boundary layer transition. Saric [19] conducted a comprehensive 69
examination of hypersonic boundary layer transition prediction, emphasizing the poten- 70
tial applicability of the eN method for predicting two-dimensional hypersonic transition 71
locations in his work [20]. Additionally, Chen et al. [21] explored boundary layer transi- 72
tion on a cone and plate at Mach number 3.5 using the eN method. There results indicate 73
that when N =10, theoretical results are in excellent agreement with the transition data 74
obtained in the experiments. Juliano et al. [22] investigated the influence of the noise en- 75
vironment on Hifire-5 elliptic cone transition in the static wind tunnel and found that N = 76
3.5 in the case of noise and N = 8 in the quiet free stream condition. Su et al. [32] improved 77
the eN method while considering the modal transformation, the results show that by ne- 78
glecting the disturbance decay between the two instability modes, the proposed strategy 79
provides more accurate results than existing strategies. LST method has also been applied 80
to other studies on boundary layer stability. Zhu et al. [23] conducted a comprehensive 81
investigation employing wind tunnel experimentation and stability analysis to explore 82
the nonlinear interplay between high-frequency and low-frequency modes on porous sur- 83
faces. Their findings revealed that the attenuation of near-wall disturbances by porous 84
surfaces induced alterations in the spatial distribution characteristics of fundamental res- 85
onance disturbances. This disruption resulted in the disintegration of phase-locked rela- 86
tionships and a concomitant suppression of the instability of fundamental oblique waves. 87
Additionally, the study highlighted the significant role of porous surfaces in the substan- 88
tial suppression of aerodynamic heating and the postponement of the transition position. 89
In a recent study, Klothakis et al. [24] employed the Direct Simulation Monte Carlo 90
(DSMC) method to computationally analyze the steady laminar flow over a semi-infinite 91
flat plate. The obtained results demonstrated a favorable agreement with corresponding 92
solutions derived from boundary layer considerations. Following this, the researchers 93
conducted a comparative examination of the linear stability characteristics between the 94
DSMC-generated base flows and those derived from classical Navier–Stokes-based pro- 95
files. The findings revealed a slight increase in stability for flows obtained through the 96
DSMC method when compared to their Navier–Stokes counterparts. 97
In this study, experimental investigations were conducted to examine hypersonic 98
boundary layer transition on a flat plate, employing Infrared (IR) thermography and PCB 99
Appl. Sci. 2023, 13, x FOR PEER REVIEW 3 of 13

technology. Additionally, linear stability analysis was employed to calculate the growth 100
rate of the second mode instability in the boundary layer across various Reynolds num- 101
bers. The primary focus of this paper centers on characterizing the transition position ob- 102
served in the experiments, along with analyzing the growth rate and amplitude of dis- 103
turbances within the hypersonic boundary layer. Furthermore, the determination of the 104
N factor at the transition location of the flat plate in the NHW hypersonic wind tunnel 105
was performed using the eN method. 106

2. Experiment facility and model 107

2. 1. Wind Tunnel 108
The experimental investigations were carried out within the Hypersonic Wind Tun- 109
nel (NHW) at Nanjing University of Aeronautics and Astronautics. This wind tunnel op- 110
erates in a blow-down and vacuum suction mode, allowing for a variable operating Mach 111
number ranging from 4 to 8 through adjustments to the corresponding Laval nozzle. The 112
wind tunnel is equipped with five sets of axisymmetric Laval nozzles, each with a 500mm 113
diameter, enabling the provision of a stable flow field at Mach number 5 for approxi- 114
mately 8 seconds during experimental procedures. Optical accessibility is facilitated by 115
two optical windows on each side of the plenum chamber, suitable for schlieren imaging. 116
Additionally, a rectangular window situated above the plenum chamber serves as an op- 117
tical access point for high-speed cameras or infrared cameras. 118

119
Figure 1. Sketch of NHW wind tunnel. 120

2. 2. Experiment Model 121

Figure 2 displays a photograph of the flat plate model positioned within the wind 122
tunnel at a zero angle of attack. The primary structure of the model is constructed from 123
stainless steel to guarantee adequate stiffness. Positioned centrally within the model is an 124
embedded polyether ether ketone (PEEK) component. PEEK is a high-temperature re- 125
sistant, high-strength resin material with excellent insulation performance. By utilizing 126
the insulation performance of PEEK material, the strength of high-speed airflow and wall 127
friction can be well reflected on the surface temperature of the model, making it conven- 128
ient for monitoring by infrared thermal imagers. In addition, PEEK also has a high infra- 129
red emissivity, making the results monitored by the infrared thermal imager highly accu- 130
rate, which makes it very suitable for the application of IR thermography. Juliano et al. 131
[22] used infrared thermal imaging technology to study the surface transition phenome- 132
non of the HIFiRE-5 model made of polyether ether ketone (PEEK) material under Ma=6 133
conditions. The experimental results showed that the infrared thermal image can clearly 134
display the high and low temperature bands flowing towards the model surface and can 135
capture the transition process. The Mach 5 free-stream direction is from left to right. The 136
infrared camera (FLIR T630sc) observes the flat plate from the upper observation hole 137
with zinc selenide (ZnSe) window. The detailed flow conditions in experiment are seeing 138
in the Table 1. The unit Reynolds number is changed by adjusting the total pressure of the 139
Appl. Sci. 2023, 13, x FOR PEER REVIEW 4 of 13

free-stream flow. Three free-stream flow conditions with different Reynolds numbers are 140
set up to study the effect of unit Reynolds number on transition Reynolds number. 141

142
Figure 2. Schematic diagram of flat plate installation in wind tunnel. 143

Table 1. Flow condition. 144

Flow condition P0/kPa T0/K Re/m/106

Case1 200.1 501 2.52
Ma=5 Case2 490.5 491 6.38
Case3 729.9 539 8.20
2. 3. Heat Flux Calculating 145
In this experiment, the heat flux distribution on the surface of the plate is obtained 146
by solving the one-dimensional inverse heat transfer equation [25]. The surface tempera- 147
ture of the model is recorded by IR image. A constant initial temperature was assumed 148
through the PEEK. According to the one-dimensional heat conduction equation: 149

T  2T
= 2 (1)
t y
where α is the thermal diffusivity, which is related to the materials. In equation (1), 150
the time term adopts the first-order forward difference, and the space term adopts the 151
second-order central difference scheme. The finite difference scheme is presented as fol- 152
low: 153

Ti +1, j − Ti , j  (Ti , j −1 − 2Ti , j + Ti , j +1 )

= (2)
t ( y)2
Δy takes the appropriate value to ensure the numerical stability. Utilizing the tem- 154
perature distribution within the PEEK model, the heat flux for individual pixels at each 155
time step is computed through the application of Fourier's law, as expressed below [25]: 156

T
q = −k (3)
y
The dimensionless Stanton number (St) is derived from the surface heat flux by uti- 157
lizing the free stream parameters. 158

q
St = (4)
U C p (T0 − Tw )
Where the subscript ∞ is represent the free stream condition. Cp, T0, Tw represent 159
the constant pressure specific heat, flow total temperature and wall temperature, respec- 160
tively. 161
2. 4. Numerical Setup 162
Appl. Sci. 2023, 13, x FOR PEER REVIEW 5 of 13

In this investigation, a two-dimensional, compressible Navier–Stokes equations are 163

solved utilizing an upwind finite-volume formulation. The transition model employed in 164
all simulations is the k-ω SST model, introduced by Menter [26] to address certain limita- 165
tions within the realm of two-equation turbulence models. The k-ω SST model integrates 166
the advantages of Wilcox’s k-ω model [27] for capturing turbulent flow near the wall and 167
the characteristics of the k-ε model in regions far from the wall. This integration is 168
achieved through a coupling function, ensuring the conservation of compressible equa- 169
tions for turbulent kinetic energy and specific dissipation rate, respectively. 170

 ( k )  ( u j k ) u j   k 
+ =  ij −   k +   ( +  k T )  (5)
t x j x j x j  x j 

And 171

 (  )  (  u j )  u    
+ =  ij j −  2 +   ( +   T ) 
t t  T x j x j  x j 
(6)
 k 
+2 (1 − F1 )  2
 x j x j
where β, β*, γ, σk, σw2 are closure constant coefficients. 172
The model is completed by the kinematic viscosity limitation of the form 173

k
T =
 F  (7)
max   , 2 
 a1 
where  is the vorticity magnitude, a constant coefficient and a function with as- 174
ymptotic behavior. Boundary conditions on a smooth wall write 175

6
k = 0, w = 10 (8)
1 y 2
with being a constant coefficient. 176
The numerical calculation model conditions are the same as the experimental condi- 177
tions, with a length of 0.5m and a height of 0.1m. Calculate the number of grids to 200 × 178
150, the height of the first layer of grid on the wall is 1 × 10-5m, satisfying y+ less than 1. 179
The schematic diagram of the grid is shown in Figure 3. The incoming flow boundary 180
condition is the pressure far field, the outlet is the pressure outlet boundary condition, 181
and the upper and front areas of the plate are set with symmetric boundary conditions. 182
0.1
y/m

0.05

0 0 0.1 0.2 0.3 0.4 0.5

x/m 183
Figure 3. Schematic diagram of flat plate mesh division and boundary conditions. 184

2. 5. Linear Stability Theory 185

Linear Stability Theory (LST) is established under the assumption of small perturba- 186
tions [4], under this assumption the perturbation can be written in the form of the sum of 187
the basic flow and the perturbation quantity: 188

p ( x, y , z , t ) = p ( x, y ) + p ' ( x, y , z , t ) (9)
Appl. Sci. 2023, 13, x FOR PEER REVIEW 6 of 13

In hypersonic boundary layer, p (x, y, z, t) representing u, v, w, ρ, T. The basic flow 189

field is obtained by solving the Blasius equation. It is assumed that the flow in the bound- 190
ary layer is parallel flow, the small perturbation p’ can be written in the form of the trav- 191
eling wave as follows: 192

p ' ( x, y, z , t ) = p ( y )ei ( x +  z −t ) + c.c. (10)

Where α, β, ω is the streamwise wave number, the spanwise wave number and the 193
frequency, respectively. In addition, c. c. represents conjugate complex number. The gen- 194
eralized eigenvalue problem can be obtained by putting Equation (2) into the wave equa- 195
tion and ignoring the second-order small quantities. The eigenvalue is the streamwise 196
growth rate, and the eigenvector represents the disturbance shape along the wall normal. 197
In spatial mode, β and ω is a real number, α is an imaginary number: 198

 =  r + i i (11)

The real part represents the wave number of the disturbance, and the imaginary part 199
represents the disturbance growth rate. When αi < 0, the disturbance is unstable along the 200
flow direction. 201
In the realm of laminar-turbulent transition (LST) analysis, a widely applied ap- 202
proach involves the computation of the N factor using a semi-empirical method known 203
as the eN method. The determination of the N value is calculated by the following formula 204
[22]. 205

x
N = −   i dx (12)
x0

Where x0 represents the flow direction position where the disturbance wave first be- 206
gins to be instability, and x0 is different at different frequencies. In general, the NT in low- 207
speed wind tunnel is generally 6~9, while in hypersonic wind tunnel, the divergence of 208
NT is quite large. In the case of hypersonic flow, the N value as the transition criterion 209
needs to be determined by wind tunnel experiments. 210
2. 6. Steady Base Flow 211
The calculation of the stability equation requires a high precision basic flow field. For 212
the flat plate boundary layer, the basic flow field can be obtained by calculating the com- 213
pressible boundary layer equation [28]. Wall temperature, Prandtl number and specific 214
heat ratio are assumed to be constant. The viscosity coefficient adopts Sutherland viscosity 215
law. 216

(  f ) + 12 ff
'' ' ''
=0

c p  fg ' +  (  g '   ) +   Ec  ( f '' ) = 0

1 1 2
(13)
2 Pr
1
=
g
In this paper, the laminar boundary layer flow should be obtained for LST analysis 217
by solving the compressible boundary layer equation. For solving the differentia equa- 218
tions, the physical quantities are nondimensionalize by the free flow parameters, and ref- 219
erence length. The dimensionless compressible boundary layer equations are as follows, 220
where Ec, k, Pr is Eckert number, Thermal conductivity and Prandtl number, respectively. 221
In this work, we consider the boundary layer of a perfect gas with Prandtl number Pr=0.71 222
and specific heat ratio γ=1.4, the viscosity law satisfies the Sutherland's law. 223
The mean flow variables of boundary layer at different unit Reynolds numbers are 224
in Figure 4. As the unit Reynolds number increases, the boundary layer thickness gradu- 225
ally diminishes. 226
Appl. Sci. 2023, 13, x FOR PEER REVIEW 7 of 13

227
Figure 4. Comparation of mean flow variables at x=0.1m. 228

3. 3. Experimental Results and Stability Analyses 229

3.1. IR Results 230

Figure 5 shows the surface temperature distribution of flat plate under different unit 231
Reynolds number. At a lower Reynolds number (Figure 5 (a)), the temperature at the lead- 232
ing edge of the flat plate is the highest and gradually decreases downstream. With the 233
increase of Reynolds number (Figure 5 (b) (c)), the plate surface shows different tempera- 234
ture distribution, and there is an obvious temperature difference along the flow direction. 235
It can be seen from the IR image that the transition region presents an irregular "wavy 236
shape", Willems et al. [29] also found this "wavy" transition non-uniformity in the study 237
of the boundary layer transition of the flat plate and analyzed that it may be caused by the 238
natural incoming flow and the disturbance non-uniformity of the forefront of the flat 239
plate. 240
The mean Stanton number distributions calculated from IR images are shown in Fig- 241
ure 6. In the case of lower unit Reynolds number (Figure 6(a)), limited by the length of the 242
plate, IR images only captures the beginning of the BL transition. With the increasing of 243
unit Reynolds number, the transition position moves up stream, whereas, the transition 244
Reynolds number is increasing, and the complete laminar to transition to turbulent pro- 245
cesses can be observed. As a comparison, the results of full laminar flow and full turbu- 246
lence are also given as a reference. That the calculation results of infrared thermal imaging 247
are in good agreement with the calculation results in the laminar flow area, and the change 248
of flow direction Stanton number is within the range of laminar flow and turbulence. 249
Appl. Sci. 2023, 13, x FOR PEER REVIEW 8 of 13

250
Figure 5. Surface temperature distribution of flat plate under different unit Reynolds number: (a) 251
Reunit = 2.56×106; (b) Reunit = 6.38×106; (c) Reunit = 8.20×106. 252

×10-4 ×10-4 ×10-4

10 10 10
laminar （a） laminar （b） 8 laminar （c）
8 8
k-ω SST k-ω SST k-ω SST
Exp_aver Exp_aver 6 Exp_aver
6 6
St St St 4
4 4

2
2 2
0.2 0.4 0.6 0.8 1 0.5 1 1.5 2 2.5 0.5 1 1.5 2 2.5 3 3.5
Rex Rex Rex 253
Figure 6. Mean Stanton number distribution of plate flow direction under different unit Reynolds 254
number: (a) Reunit = 2.56×106; (b) Reunit = 6.38×106; (c) Reunit = 8.20×106. 255

Figure 7 represents the transition Reynolds number distribution at different 256

spanwise positions under different unit Reynolds numbers. Moreover, the transition 257
Reynolds number calculated by the compressibility modified transition model γ-Reθt [30] 258
(shown as CC in the figure) is given. It can be found that for the lower unit Reynolds 259
number (Figure 7 (a)), the transition Reynolds number is also relatively small; with the 260
increase of the unit Reynolds number of the free stream, the transition Reynolds number 261
increases accordingly. In addition, it also shows that the spanwise distribution of transi- 262
tion position on the surface of the flat plate is uneven, and the spanwise distribution of 263
transition Reynolds number increases with the increase of unit Reynolds number, as same 264
as Figure 5. In this paper, the unit Reynolds number has a great influence of transition 265
Reynolds number. Whereas some wind tunnel experiments show the opposite results. A 266
more in-depth investigation is warranted to explore the impact of unit Reynolds number 267
on the transition Reynolds number. 268
Appl. Sci. 2023, 13, x FOR PEER REVIEW 9 of 13

Ma5case1 Retr (a) Ma5case2 Retr (b) Ma5case3 Retr (c)

cc 1.4 cc 1.4 cc 1.4

1.2 1.2 1.2

1.0 1.0 1.0

0.8 0.8 0.8

0.6 0.6 0.6

-60 -40 -20 0 20 40 60 -60 -40 -20 0 20 40 60 -60 -40 -20 0 20 40 60
y/mm y/mm y/mm 269
Figure 7. Transition Reynolds number distribution of plate spanwise at different unit Reynolds 270
numbers: (a) Reunit = 2.56×106; (b) Reunit = 6.38×106; (c) Reunit = 8.20×106. 271

3.2. Instability Characteristics Along Streamwise Direction 272

The inhouse LST code written by FORTRAN used in this paper is shown in Ref [31]. 273
Figure 8 illustrates the variations in disturbance growth rates within the ω-β plane at 274
Rex=6.5×105 under the free-stream conditions of case 3. Two distinct modes coexist in the 275
boundary layer— the lower frequency corresponds to the first mode, while the higher 276
frequency characterizes the second mode. Notably, the bandwidth of the first mode sur- 277
passes that of the second mode. The diagram emphasizes the prevalence of the second 278
mode in the hypersonic boundary layer, notably featuring a predominantly two-dimen- 279
sional disturbance. Conversely, the first mode, exhibiting the highest amplification, pre- 280
dominantly occurs near β=0.8, signifying its three-dimensional perturbation nature. 281

282
Figure 8. Growth rate in ω-β plane: (a) Reunit = 6.38×10 ; (b) Reunit = 8.20×10 .
6 6 283

The evolution of instability waves along the streamwise direction is obtained and 284
analyzed. Figure 9 shows the PSDs of boundary-layer instability waves at several stream- 285
wise positions under different unit Reynolds number by PCB. 286

287
Figure 9. Comparison of PCB measured spectra along the streamwise direction: (a) Reunit = 6.38×106; 288
(b) Reunit = 8.20×106. 289
Appl. Sci. 2023, 13, x FOR PEER REVIEW 10 of 13

As illustrated in Figure 9, for a unit Reynolds number (Reunit) of 6.38×106, the absence 290
of a discernible peak characterizing second-mode waves in the Power Spectral Density 291
(PSD) before the streamwise location of x=180mm indicates that second-mode disturb- 292
ances have not yet developed, signifying a laminar boundary layer state at this location. 293
However, a subtle peak with broadband characteristics around 100 kHz emerges in the 294
PSD at x=240 mm, suggesting the initiation and growth of second-mode waves. Subse- 295
quently, the PSD of the second mode remains consistent between x=330mm and x=360mm, 296
indicating saturation of the second-mode disturbances. In the case of a higher unit Reyn- 297
olds number, Reunit =8.20×106, the peak in the PSD at the streamwise location of x=240 mm 298
surpasses that of Reunit =6.38×106, indicating the earlier appearance and growth of second- 299
mode disturbances before x=240 mm. Beyond x=330 mm, the energy associated with sec- 300
ond-mode waves exhibits a gradual increase and disperses into neighboring frequency 301
domains, signaling the imminent breakdown of second-mode waves and the transition of 302
the boundary layer. 303
According to previous studies, when the Ma>4, the second mode wave is more un- 304
stable, always leading to transition. Figure 10 shows the evolution of the growth rate of 305
second mode disturbances in different frequencies along the flow direction by LST. The 306
second modes at different frequencies start to grow from different stations, and the high 307
frequency modes usually begin to arise and grow at first. Figure 10 (b) indicates that, at 308
the same unit Reynolds number condition, the higher frequency of the second mode wave 309
can achieve higher growth rate than those low frequency. However, due to the attenuation 310
of the high-frequency disturbance, it develops to a very short distance in the direction of 311
flow, as shown in Figure 11. As the high-frequency modes begin to decay, the low-fre- 312
quency modes in the boundary layer begin to appear and grow along the flow direction. 313
Although the growth rate of lower frequency disturbance is smaller than the higher fre- 314
quency, whereas it usually becomes the key to promote transition because of its larger 315
propagation distance in the boundary layer. 316
0.08 0.08
(a) (b)
190kHz 190kHz
152kHz 152kHz
129kHz 129kHz
0.06 0.06
114kHz
−i

−i

0.04 0.04

0.02 0.02

0.00 0.00
0.05 0.10 0.15 0.20 0.25 0.30 0.35 0.40 0.45 0.50 0.05 0.10 0.15 0.20 0.25 0.30 0.35 0.40 0.45 0.50
x/m x/m
317
Figure 10. The growth rate of second mode waves along flow directions predicted by LST: (a) Reunit 318
= 6.38×106; (b) Reunit = 8.20×106. 319

As is shown in Figure 9, with the increase of unit Reynolds number, the initial posi- 320
tion of disturbance growth at the same frequency gradually moves backward. Under the 321
condition of Reunit =2.52×106(Figure 9(a)), the second mode of 190kHz in the boundary layer 322
has decreased. Moreover, the growth rate of the disturbance is higher than that in the case 323
of low unit Reynolds number and moves further forward. 324

3.3. eN method 325

The distribution of N-factor of a range of representative frequency disturbances along 326
the flow direction are shown in Figure 11. The N-factor is sensitive to frequency and free- 327
stream unit Reynolds number. Because of the larger propagation distance in the boundary 328
layer of relatively low-frequency disturbance, it is easier to reach a larger N value, as sug- 329
gested in Figure 10. Because the transition position of the plate in the experiment is obvi- 330
ously “wavy”, the transition position calculated by the transition model γ-Reθt is used as 331
the ‘average’ transition position of the flat plate in the experiment. The red dashed line in 332
Figure 10 represents the transition position obtained from the heat flow in the experiment. 333
Appl. Sci. 2023, 13, x FOR PEER REVIEW 11 of 13

The N-factor of the transition location of two case of unit Reynolds number are 0.98 and 334
1.25, respectively. Figure 12 illustrates eigenfunction of two cases in different unit Reyn- 335
olds numbers at x=130mm. There is no significant difference between the shape function 336
of flow velocity u and temperature under the two experiments conditions. The normal 337
distance of disturbance decreases with the increase of unit Reynolds number. 338

(a) 2.00
XT_Exp
1.75
1.50

1.25 f=190kHz
f=182kHz
f=175kHz
N

1.00 f=167kHz
f=160kHz
0.75 f=152kHz
f=145kHz
f=137kHz
0.50 f=130kHz
f=120kHz
0.25 f=113kHz

0.00
0.1 0.2 0.3 0.4
3.0
(b)
XT_Exp
2.5

2.0 f=227kHz
f=220kHz
N

f=213kHz
1.5 f=206kHz
f=198kHz
f=190kHz
1.0 f=183kHz
f=175kHz
f=167kHz
0.5 f=160kHz
f=153kHz
0.0
0.1 0.2 0.3 0.4
x/m
339
Figure 11. N-factor predicted by LST.: (a) Reunit =6.38×106; (b) Reunit = 8.20×106. 340

(a) (b)
0.125 1.2

case2 1.0
case2
0.100 case3 case3

0.8
0.075
|T'|
|u'|

0.6

0.050
0.4

0.025
0.2

0 1 2 3 4 5 0 1 2 3 4 5
h h
341
Figure 12. Eigenfunction predicted by LST at x=130mm. 342

4. Discussion 343
This study systematically examined the instability characteristics of the hypersonic 344
boundary layer over a flat plate at a zero angle of attack, employing a combined approach 345
of experimental analysis and linear stability analysis. The application of PCB technology 346
and infrared thermography facilitated the characterization of instability waves and ena- 347
bled the calculation of the Stanton number along the streamwise direction. Additionally, 348
Appl. Sci. 2023, 13, x FOR PEER REVIEW 12 of 13

the eN method was implemented to determine the N factor at the transition location within 349
the wind tunnel. The conclusions derived from this study are outlined as follows: 350
(1) The transition region presents an irregular "wavy shape" in the wind tunnel 351
results; it may be caused by the natural free stream and the disturbance non-uniformity 352
of the front edge of the flat plate. Moreover, the unit Reynolds number of free stream has 353
a great effect of the transition Reynolds number of flat plate boundary layer. As the unit 354
Reynolds number increases, the transition position moves forward, and the transition 355
Reynolds number also increases. 356
(2) The LST results show that the first mode and the second mode are both present 357
in the hypersonic boundary layer at the Mach number 5. In accordance with the findings 358
from Linear Stability Theory (LST), the initial mode wave, manifesting three-dimensional 359
disturbance characteristics, demonstrates the highest amplification rate. Concurrently, the 360
second-mode wave in a two-dimensional configuration with β = 0 emerges as the epitome 361
of instability. For the second modes, there is an augmentation in the maximum frequency 362
growth rate within the identical frequency mode, accompanied by a greater propagation 363
distance along the direction of free flow, corresponding to the increment in unit Reynolds 364
number. 365
(3) The N-factor of wind tunnel transition location predicted by LST is about 0.98 366
and 1.25 for Reunit =6.38×106 and 8.20×106, respectively. With the increase of the unit Reyn- 367
olds number, although the transition position moves forward, the N value of the transition 368
position increases due to the increase of the magnification of the disturbance. 369
370

Author Contributions: Conceptualization, Hao Dong; Validation, Shicheng Liu; Formal analysis, 371
Yinglei Jiang; Investigation, Shicheng Liu; Data curation, Yinglei Jiang; Writing – original draft, 372
Yanxin Yin; Funding acquisition, Hao Dong. 373

374

Funding: This research was funded by the National Numerical Wind-Tunnel (No. 2021-ZTNNW- 375
QT00-50042, No.2020-DY01-001), the National Natural Science Foundation of China (No. 11872208). 376

References 377
1. Marineau E C, Grossir G, Wagner A, et al. Analysis of second-mode amplitudes on sharp cones in hypersonic wind tunnels[J]. 378
Journal of Spacecraft and Rockets, 2019, 56(2): 307-318. 379
2. CHENG J Q, TU G H, ZHANG Y F, et al. Hypersonic boundary layer transition: what we know where shall we go [J]. Acta 380
Aerodynamic a Sinica, 2017, 35(3): 311-337. 381
3. Quintanilha H, Paredes P, Hanifi A, et al. Transient growth analysis of hypersonic flow over an elliptic cone[J]. Journal of Fluid 382
Mechanics, 2022, 935: A40. 383
4. Mack L M. Boundary-layer linear stability theory[R]. California Inst of Tech Pasadena Jet Propulsion Lab, 1984. 384
5. Mack L M. Linear stability theory and the problem of supersonic boundary-layer transition[J]. AIAA journal, 1975, 13(3): 278- 385
289. 386
6. Laurence S J, Wagner A, Hannemann K. Experimental study of second-mode instability growth and breakdown in a hy-personic 387
boundary layer using high-speed schlieren visualization[J]. Journal of Fluid Mechanics, 2016, 797: 471-503. 388
7. Estorf M, Radespiel R, Schneider S, et al. Surface-pressure measurements of second-mode instability in quiet hypersonic 389
flow[C]//46th AIAA Aerospace Sciences Meeting and Exhibit. 2008: 1153. 390
8. Wendt V, Simen M, Hanifi A. An experimental and theoretical investigation of instabilities in hypersonic flat plate boundary 391
layer flow[J]. Physics of Fluids, 1995, 7(4): 877-887. 392
9. Zhao J S，Liu S，Zhao L，et al. Numerical study of total temperature effect on hypersonic boundary layer transition［J］. Physics of Fluids 393
，2019，31（11）:114105. 394
10. Zhang C H, Tang Q, Lee C B. Hypersonic boundary-layer transition on a flared cone[J]. Acta Mechanica Sinica, 2013, 29(1): 48- 395
54. 396
11. Zhu Y, Zhang C, Chen X, et al. Transition in hypersonic boundary layers: Role of dilatational waves[J]. AIAA Journal, 2016, 397
54(10): 3039-3049. 398
12. Thele M, Selcan C, Sander T, et al. Bluntness-Dependent Hypersonic Boundary-Layer Modes’ Excitation[J]. Journal of Spacecraft 399
and Rockets, 2022, 59(5): 1613-1622. 400
13. Borg M P, Kimmel R L. Ground test of transition for HIFiRE-5b at flight-relevant attitudes[J]. Journal of Spacecraft and Rockets, 401
2018, 55(6): 1329-1340. 402
Appl. Sci. 2023, 13, x FOR PEER REVIEW 13 of 13

14. Zhao Y F, Liu W, Gang D D, et al. Study of surface roughness induced supersonic boundary layer transition[J]. Journal of 403
Astronautics, 2015, 36(6): 739-746. 404
15. Zhu Y, Zhang C, Chen X, et al. Transition in hypersonic boundary layers: Role of dilatational waves[J]. AIAA Journal, 2016, 405
54(10): 3039-3049. 406
16. Hao D, Shicheng L I U, Xi G, et al. Numerical and experimental investigation into hypersonic boundary layer transition induced 407
by roughness elements[J]. Chinese Journal of Aeronautics, 2019, 32(3): 559-567. 408
17. Nakagawa K, Tsukahara T, Ishida T. DNS Study on Turbulent Transition Induced by an Interaction between Freestream Tur- 409
bulence and Cylindrical Roughness in Swept Flat-Plate Boundary Layer[J]. Aerospace, 2023, 10(2): 128. 410
18. Egorov I V, Novikov A V. Direct numerical simulation of laminar–turbulent flow over a flat plate at hypersonic flow speeds[J]. 411
Computational Mathematics and Mathematical Physics, 2016, 56(6): 1048-1064. 412
19. Saric W, Reshotko E, Ama l D, Hypersonic laminar-turbulent transition, AGARD, 1998, AR-3l9. 413
20. Mason, W.H, Fundamental Issues in Subsonic/ Transonic Expansion Corner Aerodynamics, AIAA Paper, 1993, 0649. 414
21. Chen F J, Malik M R, Beckwith I E. Boundary-layer transition on a cone and flat plate at Mach 3.5[J]. AIAA journal, 1989, 27(6): 415
687-693. 416
22. Juliano T J, Paquin L, Borg M P. Measurement of HIFiRE-5 boundary-layer transition in a Mach-6 quiet tunnel with infrared 417
thermography[C] 54th AIAA Aerospace Sciences Meeting. 2016: 0595. 418
23. Zhu W, Shi M, Zhu Y, et al. Experimental study of hypersonic boundary layer transition on a permeable wall of a flared cone[J]. 419
Physics of Fluids, 2020, 32(1). 420
24. Klothakis A, Quintanilha Jr H, Sawant SS, Protopapadakis E, Theofilis V, Levin DA. Linear stability analysis of hypersonic 421
boundary layers computed by a kinetic approach: A semi-infinite flat plate at 4.5≤ M∞≤ 9. Theoretical and Computational Fluid 422
Dynamics. 2022 Feb;36(1):117-39. 423
25. Boyd C F and Howell A. Numerical investigation of one-dimensional heat-flux calculations[R]. Technical Report NSWCDD/TR- 424
94/114, Dahlgren Division Naval Surface Warfare Center, Silver Spring, MD 20903-5640, October 1994. 425
26. Menter F. R. 1994. Two-equation eddy-viscosity turbulence models for engineering applications. AIAA journal. 32(8) :1598– 426
1605. 427
27. Chedevergne F. A double-averaged Navier-Stokes k–ω turbulence model for wall flows over rough surfaces with heat trans- 428
fer[J]. Journal of Turbulence, 2021, 22(11): 713-734. 429
28. White, F. M., Viscous fluid flow, 3rd ed., McGraw-Hill series in mechanical engineering, McGraw-Hill Higher Education, New 430
York, NY, 2006. 431
29. Willems S, Gülhan A, Steelant J. Experiments on the effect of laminar–turbulent transition on the SWBLI in H2K at Mach 6[J]. 432
Experiments in Fluids, 2015, 56(3): 49. 433
30. Menter F R, Langtry R B, Likki S R, et al. A correlation-based transition model using local variables—part I: model formula- 434
tion[J]. Journal of turbomachinery, 2006, 128(3): 413-422. 435
31. Guo X, Tang D, Shen Q. Boundary layer stability with multiple modes in hypersonic flows[J]. Modern Physics Letters B, 2009, 436
23(03): 321-324. 437
32. Tao S, Su C, Huang Z. Improvement of the e N method for predicting hypersonic boundary-layer transition in case of modal 438
exchange[J]. Acta Mechanica Sinica, 2023, 39(6): 122416. 439

Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual au- 440
thor(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to 441
people or property resulting from any ideas, methods, instructions or products referred to in the content. 442