Boundary Layer Separation Control in Y-intake duct

Adapter with Different Degree Roughness

Abstract-despite the fact that enhancing duct performance is a the flow can be reversed half of the time upstream of the
primary goal in device design, there are various elements that detachment point while maintaining a positive average
influence the device's efficiency during operation. The effects velocity. In the rough-wall example, the separated shear
of a wide range of surface roughness on stream flow and power layer has a greater turbulent kinetic energy (TKE). The TKE
loss inside a Y-intake duct adapter are investigated in this
research. On a Y-intake Duct Adapter, a complete program of
grows faster there than it does at the separation point, and
2D Computational Fluid Dynamics (CFD) modeling, as well as the peak TKE occurs near the separation point. The
an extensive range of experiments, were carried out to quantify roughness-induced momentum deficit plays a key role in
performance degradation owing to surface roughness. Surface these alterations once again. As shown in Fig. 1,[7] flow
roughness inside the duct has a crucial role in boundary layer control can be divided into two branches: turbulence
separation, according to the findings. Surface roughness manipulation and laminar flow. In addition, there are two
increases turbulence and lowers the active fluid energy types of turbulence manipulation: active and passive
required for duct performance, lowering the duct's turbulent flow management. The goal of low control using
performance. Surface roughness has a negative impact on the
sandpaper over a wind turbine airfoil [8] was to see how the
drag coefficient, pressure coefficient, which is evaluated. It is
proven that a duct with a roughness height of 375 μm has a roughness of the sandpaper affected the low Reynolds flow
pressure coefficient of 20% greater than a smooth duct wall over the NACA 4412 airfoil. To determine where sandpaper
(zero roughness height), and the greatest percentage of Error should be placed on the airfoil and save money on
Calculation for the Validated Cases might be up to experiments, First, a numerical analysis was carried out at
7.22%Ks=375μm)and 6.98% (Ks=345 μm) and 7.028% low Reynolds numbers ranging from 2.5×10 4 to 1×10 5.
(Ks=250 μm) and 6.98% for case without roughness.
The position of sandpaper was chosen as 15%- 25% chord
Keywords- Y-intake duct adapter; surface roughness; boundary length as a result of the numerical analysis, and roughness
layer separation; pressure coefficient. height (k) was also chosen as 0.5mm and 1mm (k is
normalized as k/c throughout the rest of this study). The
I. INTRODUCTION impact of surface roughness on turbulent boundary layers
Separation of the turbulent boundary layer is one of the has been investigated by many researchers. Acharya et al.
most important processes affecting the efficiency of flow [9] and Brzek et al. [10] showed that surface roughness
devices such as airplane wings, turbine and compressor increases the friction coefficient (Cf) in zero pressure
blades, and curved ducts; their greatest efficiency is gradient (ZPG) boundary layers. For a rough surface, the
generally around the commencement of separation. In these total drag was composed of the pressure drag on roughness
arrangements, reliable prediction of incipient detachment elements and the friction drag on the surface [11,12]. The
and reattachment is crucial. During operating under high growth rates of the displacement thickness (d) and
pressure and temperature settings, surface roughness momentum thickness (h) increased as Cf was increased [9].
increases dramatically due to erosion, corrosion, and The pipe flow study of Nikuradse [1] and the ZPG plate
deposition. As a result, surface roughness has an impact on boundary layer studies of Brzek et al. [10], Krogstad et al.
duct efficiency, and the predicted life can be shortened. [13], and Krogstad and Antonia [14] all showed that
Since the 1930s, roughness has been tested [1-2]. The roughness increases the normalized mean velocity defect (1-
majority of the research, on the other hand, has concentrated u
). Attention has also been paid to the favorable pressure
on equilibrium turbulent flows (flat-plate boundary layers, U∞
channels, pipes). The interaction of roughness and pressure gradient (FPG) effects on the smooth plate turbulent
gradients, particularly unfavorable pressure gradients (APG) boundary layers [15]. The FPG decreased the mean velocity
that produce separation, is poorly understood. Roughness defect [16,17] and shape factor [16,18]. Joshi et al. [15] and
reduces the drag of bluff bodies at low Reynolds numbers Escudier et al. [18] showed that the FPG increases near the
by producing early transition of the flow and so delaying φ
separation, according to every textbook. In fully turbulent surface (y/δ <0.1 in Joshi et al. [15] and <2000 in
flows, roughness has the opposite effect: separation occurs v
sooner when roughness is present [3-4]. In comparison to Escudier et al. [16]) and decreases it in the outer region.
smooth-wall boundary layers, how separation happens in According to a survey of the literature, there is a gap in our
rough-wall boundary layers is not well understood, and understanding of the impact of surface roughness on duct
many problems remain unresolved. Separating turbulent performance. A detailed CFD simulation and a series of
boundary layers over smooth and rough flat plates are laboratory experiments for a Y-intake duct were carried out
studied by large-eddy simulations [5-6], The roughness in this study, covering a wide range of roughness values. to
element also impacts the near-wall flow's intermittency, thus determine the effects of roughness on the performance of
the Y-intake duct adapter. The major purpose of this work is
to investigate the negative impacts of roughness on a duct The choice of a proper computational domain is a
using both numerical and experimental methods. The critical aspect in ensuring that the CFD model is
numerical and experimental model results are presented and representational of reality [19]. The domain should be
analyzed in the following sections. adjusted to reduce computational run times while
simultaneously providing enough space for proper meshing.
Furthermore, the CFD domain should be suitable for
reproducing the duct without the effects of the wall
roughness. To analyze the performance of the duct, a 2D
transient time-accurate and dynamic mesh CFD model was
developed using the COMSOL Multiphysics 5.5. geometry
of the duct used that shown in Figure 3. was design in
COMSOL. Mesh generation was done in the COMSOL
CFD software as shown in figure 4. The angle in the duct
was modelled with an unstructured mesh, whereas the rest
of the domain was modelled with a structured mesh. Table 2
Fig. 1 Various type of flow control utilization [7]. shows the total number of cells in this simulation (after grid
sensitivity study). To simulate the problem, the RNG K-E
II. NUMERICAL MODELING turbulence model was employed, and the speed and pressure
equations were coupled. At the entrance, the air velocity
A. Model Geometry was 4.5 m/s. The velocity intake was employed as the entry
Some key parameters of Y-intake duct adapter are condition, while the pressure outlet was used as the exit
presented in Table 1. The air velocity was set at 4.5 m/s and condition. The wall boundary condition was applied to the
height of height and length of the duct were selected as 0.9 domain's bottom, top, and side bounds. The simulation was
and 0.26 m, respectively. run in transient mode, with a time step of 0.001s chosen
after a time dependence test.
TABLE I

Duct geometrical and stream specifications.

Height of the duct 0.15 m

Length of the duct 0.9 m
Width of the duct 0.26 m
Type of duct Y-intake duct adapter
Velocity of air 4.5 m/s

Fig. 2 depicts the design of a Y-intake duct adapter in

relation to air flow. The interaction of the Y-intake duct
with the air is complicated, and a number of parameters are
required to fully analyze it.

Fig. 2 Computational domain.

Fig. 2 Schematic of the Y-intake duct adapter.

 G2 6494 390 150 11.2 0.714 16.2 4.32

G3 17700 712 378 11.3 0.884 16.6 2.4

G4 19076 714 414 11.3 0.0 16.62 0.12

1
G5 28696 736 632 11.3 0.0 16.62 0.0
1

III. EXPERIMENTAL SETUP

The separation of the duct's boundary layer was
measured and examined using an experimental setup that
was designed and manufactured (Figure 5). A Perspex Y-
intake duct adaptor with an air blower and honey comb was
created. A number of holes were drilled into the Y-intake
duct for the purpose of measuring velocity. The distance
between each hole was determined, and the position of these
Fig.4 domain and intensive mesh on walls of the Y-intake duct adapter holes was established in the duct's center. A pitot-static tube
(using COMSOL Multiphysics)
linked to a manometer was used to measure the pressure
change (fig. 6 (a)).
The air blower's velocity can be adjusted. A model and
a prototype of the Y-intake duct were created (fig. 7 (b)) (in
order to see the areas where the separation occurs by the
TABLE II smoke tunnel). To assess losses owing to surface roughness,
a comprehensive set of laboratory tests was devised and
Grid independent test for Velocity and Roughness rate on surface [20].
carried out. The studies were conducted with r sandpapers
Gri Domai Bounda Ti U exit |Error|%
U at 50 cm |Error|% of consistent roughness. The effects of surface roughness on
d n ry me flow and separation were investigated using three different
elemen element (sec
ts s )
sandpapers with roughness values of (250, 345, 375, 500)
G1 2332 206 73 11.1 - 15.5 - μm. In the following sections, the experimental results will
2 be given and compared to the CFD results.

Fig.5 Measurement system for the Y-intake duct adapter.

 a. The Experimental Test Rig

b. The model inside the smoke tunnel

Fig. 6 Experiments in fluid laboratory of University of Babylon.

IV. VALIDATION OF THE CFD MODEL performance of the Y-intake duct adapter reported in the
literature
To test the validity of the CFD model, the experimental
results and outcomes from some other studies from the
literature will be compared to the results of the CFD
simulations of the Y-intake duct in this section. Pressure
Coefficient (Cp) was determined both experimentally and
computationally along the length of a Y-shaped intake duct
for this purpose [21] can be used to compute the pressure
coefficient:

P s−P si
C PR= 1 2 (1)
U
2 avi
Figure 7 depicts the computed Cp values at various
.
locations in the duct. The numerical results are in good
agreement with the experimental results, as can be shown. Fig. 7 Pressure coefficient (CP) vs. length of Y-shaped intake duct for
The overall trend of the current experimental and CFD experimental validation of the computational fluid dynamics (CFD) model.
results is also comparable with some results on the
 V. ROUGHNESS MODELING USING THE MODIFIED LAW OF k +s ¿¿, is the roughness height in viscous units,
THE WALL

The RNG model was developed using Re-Normalisation 1

ρ C μ4 √k
Group (RNG) methods [22]. It also takes the effective + ¿= k s¿ (6)
μ
viscosity in to account, meaning that another analytically ks
derived differential formula accounts for low-Reynolds-
number effects. However, an appropriate treatment of the
near-wall region has to be defined in order for this effect to
be accurately accounted for turbulent model [23]. The RNG
VI. RESULTS AND DISCUSSION
model’s transport equations which are used in the Fluent
code are as follow [24]: B. Numerical Results
B. 1 Surface Roughness Effects on Flow Pattern

μ ∂k The Y-intake duct adapter was simulated using the CFD

∂
∂t
(2)
( ρk ) +
∂
∂ xi
( ρk u i) =
∂
∂ xi [( ) ]
μ+ t
σ k ∂ xi
+ Pk −ρϵ model in six different average roughness heights (Ks) of 0,
250, 345, 375 μm at velocity 4.5 m/s. The turbulence model
should be established and validated, which is one of the
most significant components in roughness simulation. The
RNG K-ε model was utilized as the turbulence model, and
its capabilities for roughness impacts were proven by Sagol
μ ∂ϵ ϵ 2 et al. [26] and Villalpand et al. [27]. The distribution of
∂
∂t
(3)
( ρϵ ) +
∂
∂ xi
( ρϵ ui ) =
∂
∂ xi [( ) ]
μ+ t
σϵ ∂ xi
ϵ ¿
+C 1 ϵ Pk −C2 ϵ ρ
k k velocity and velocity vectors can be useful in analyzing and
understanding the phenomenon of flow over a structure such
as Y-intake ducts. Figures 8 show the results of the two-
dimensional CFD analysis. The velocity distribution is
depicted as a cross-section to better comprehend the effect
The physics interfaces Wall and Interior Wall have an of roughness height on the flow . We select a reference
option to apply wall roughness by modifying the wall angle (45) and then simulate the duct setup with multiple
functions. Cebeci [25] suggested a model which adjusts the roughness heights to provide a comparison between the duct
friction velocity for surface roughness. with different roughness heights. The duct with varying
surface roughness heights can cause turbulence in the
|u| stream flow, as can be shown. This turbulence can cause the
uτ = boundary layer to separate and the stream flow to change as
1 (4)
ln σ +w¿+ B−∆ B ¿ a result of the turbulence. Fluid turbulence constantly lowers
kv the kinetic energy, which lowers the overall duct efficiency.
The effects of roughness are insignificant at lower
roughness heights (less than 250 μm), as may be observed
Where: (see Figures 8). The velocity vectors of the fluid flow
undergo a substantial leap at high roughness, which has a
significant impact on the fluid flow (particularly at
roughness heights more than 375 μm). In the next sections,
this effect will be described and numerically measured.
 Fig. 8 velocity contours plots for 45 ° intake models with different roughness rate.

B. 2 Effects of skin friction coefficient (Cf) Friction Factor boundary layer theory to produce the data in the figure
below. the skin-friction coefficient, Cf, is shown as a
The skin friction coefficients are sometimes based on function of Reynolds number, Re. Results from the present
experimental data for duct with various amounts of study show that the skin-friction coefficient decreases as the
roughness. In the present method, experimental results for Reynolds number, Re, increases and it's also effected by
turbulent duct are fit and combined with basic flow degree of roughness shown in figure (9).

5.20E-03
A=1e-12 m

5.00E-03 A=0.001 m
A=0.002 m
4.80E-03

4.60E-03
Cf

4.40E-03

4.20E-03

4.00E-03
100000 150000 200000 250000 300000 350000
Reynolds Number

Fig. 9 Relation between skin friction coefficient and Reynold Number

At high Reynolds number, the friction factor of rough duct by the Reynold Number and roughness, and it is clearly
becomes constant, dependent only on the duct roughness. affected, from the experimental work found that the friction
For smooth duct, shown that the friction factor influenced
factor decreases as the Reynold Number increase as shown in figure 10.

Fig. 10 shows relation between Friction Factor and Reynold Number

The results of the laboratory experiments and numerical

on the effects of surface roughness on the flow. Figure 11
shows the Y-intake duct adapter with three different surface
roughness values created by sandpapers with different
surface roughness. The results are shown in Figure 11 show
that the region of separation increases nonlinearly with
increase in the surface roughness.
C. Experimental with Numerical Results

Ks: 375 μm
 Ks: 345 μm

Without sandpaper
Fig. 8 Experimental and Numerical cases with different roughness values.

VII. VALIDATION Analysis is performed for different for the mean velocity
streamlines, the streamwise mean velocity profiles, the
D. 1 Comparison of streamlines between (S. Thangam and streamwise turbulence intensity profiles and the turbulence
C.G. Spezialet (1991) [29], and present study. shear stress profiles.

Study Stream lines

Thangam and
Speziale [79]
 Present study

Fig. 12 Validation of boundary layer thickness of present work vs. S. Thangam and C.G. Spezialet (1991) [29]

D. 2 Validation between experimental and numerical study

Figure 13. show the validation between the experimental
results and the numerical results (CFD) inside the Y-intake
duct adapter with different degree of surface roughness. The
validation of the CFD data was carried out with different
velocities 4.5 (m / s), 5.5 (m / s), 6.5 (m / s), 7.5 (m / s) and
different roughness height values (0, 250 μm, 345
μm, 375 μm). A comparison of general data shows a
consensus between numerical and experimental results.

Fig. 13 Validation of the velocity profile at different roughness Height.

VII. CONCLUSIONS
An experimental and CFD examination of the impact of
surface roughness on the performance of a Y-intake duct
adapter was reported in this paper. It was discovered that
surface roughness is an important aspect to consider when [4] AUBERTINE, C. D., EATON, J. K. & SONG, S. 2004
constructing a duct, and that it should be regarded as an Parameters controlling roughness effects in a separating
early design parameter along with duct size, air velocity, boundary layer. Intl J. Heat Fluid Flow 25 (3), 444–450.
and duct shape. The duct performance decreases as the
roughness of the Y-intake duct increases. The impacts of [5] Wu, W., & Piomelli, U. (2018). Effects of surface
roughness on stream flow and duct performance were roughness on a separating turbulent boundary layer. Journal
quantified using CFD models. Experiments were also of Fluid Mechanics, 841, 552–580.
conducted on physical models of Y-intake with four
different roughness heights. The following are some of the [6] Husain, N. T., Hara, T., Buckley, M. P., Yousefi, K.,
inferences that may be derived from the findings of this Veron, F., & Sullivan, P. P. (2019). Boundary layer
study: turbulenceover surface waves in a strongly forced
condition: LES and observation. Journal of Physical
(1) Velocity distributions show that turbulence will increase Oceanography, 49(8), 1997–2015.
with increasing the surface roughness of Y-intake duct.
Turbulence reduces the kinetic energy which in turn, [7] Serdar GENÇ, M., KOCA, K., & AÇIKEL, H. H.
reduces the total duct efficiency. (2019). Investigation of pre-stall flow control on wind
turbine blade airfoil using roughness element. Energy,
(2) The surface roughness degrades the margin of pressure 176, 320–334.
coefficient, especially at high values of roughness. https://doi.org/10.1016/j.energy.2019.03.179
(3) The pressure cofficient increases as the roughness [8] Koca K. The flow control with roughness devices over
decrease, the pressure coefficient is very sensitive to the wind turbine airfoil. MSc Thesis. Kayseri, Turkey: Graduate
variations of Ks. The pressure coefficient of roughed duct School of Natural and Applied Sciences, Erciyes University;
(with Ks = 375 μm) studied in this paper was 20% higher 2016.
than the smooth duct.
4) At high intake velocities, the negative impact of surface [9] Acharya, M., Bornstein, J., and Escudier, M. P., 1986,
roughness is substantially stronger (i.e., high Reynold “Turbulent Boundary Layers on Rough Surfaces,” Exp.
numbers). Except for the air velocity, all parameters in the Fluids, 4, pp. 33–47.
Re equation were constant in the CFD simulation.
Roughness has been found to have substantially bigger [10] Brzek, B., Cal, R. B., Johansson, T. G., and Castillo,
detrimental effects at high Reynolds numbers. L., 2007, “Inner and Outer Scalings in Rough Surface Zero
Pressure Gradient Turbulent Boundary Layers,” Phys.
(5) The duct's performance is affected by an increase in Fluids, 19, p. 065101.
separation caused by surface roughness and the greatest
percentage of Error Calculation For the Validated Cases [11] Furuya, Y., Miyata, M., and Fujita, H., 1976,
might be up to 7.22 % (Ks=375 μm ) and 6.98 % “Turbulent Boundary Layer and Flow Resistance on Plates
(Ks=345 μm) and 7.028 % ( Ks=250 μm )∧6.98 % for Roughened by Wires,” ASME J. Fluids Eng., 98, pp. 635–
643.
case without roughness.
(6) The results of this study will help manufacturers and [12] Leonardi, S., Orlandi, P., and Antonia, R. A., 2007,
operators of Y-intake duct adapters in making decisions on “Properties of d- and k-Type Roughness in a Turbulent
practical steps (e.g., duct coating or cleaning) to reduce duct Channel Flow,” Phys. Fluids, 19, p. 125101.
roughness.
[13] Krogstad, P. -A° ., Antonia, R. A., and Browne, W. B.,
REFERENCES 1992, “Comparison Between Rough- and Smooth-Wall
[1] NIKURADSE, J. 1933 Laws of flow in rough pipes (in Turbulent Boundary Layers,” J. Fluid Mech., 245, pp. 599–
German). VDI Forsch. 361 (translation in NACA Tech. 617.
Rep. 1292, 1950).
[2] COLEBROOK, C. F. 1939 Turbulent flow in pipes, with [14] Krogstad, P.- A ° . and Antonia, R. A., 1999, “Surface
particular reference to the transition region between the Roughness Effects in Turbulent Boundary Layers,” Exp.
smooth and rough pipe laws. J. Inst. Civil Engrs 11 (4), Fluids, 27, pp. 450–460.
133–156.
[15] Joshi, P., Liu, X., and Katz, J., 2011, “Turbulence in
[3] SONG, S. & EATON, J. K. 2002 The effects of wall Accelerating Boundary Layers,” ASME-JSME-KSME Joint
roughness on the separated flow over a smoothly contoured Fluids Engineering Conference 2011, Hamamatsu, Japan,
ramp. Exp. Fluid 22, 38–46. July 24–29.
[16] Escudier, M. P., Abdel-Hameed, A., Johnson, M. W.,
and Sutcliffe, C. J., 1998, “Laminarisation and Re- [24] William, T. V. (2017). CFD Module User ’s Guide.
Transition of a Turbulent Boundary Layer Subjected to CFD Module User’s Guide, 1–710.
Favorable Pressure Gradient,” Exp. Fluids, 25, pp. 491–502. https://doc.comsol.com/5.3/doc/com.comsol.help.cfd/
CFDModuleUsersGuide.pdf
[17] Herring, H. J. and Norbury, J. F., 1967, “Some
Experiments on Equilibrium Turbulent Boundary Layers in [25] T. Cebeci, Analysis of Turbulent Flows, 2nd ed.,
Favorable Pressure Gradients,” J. Fluid Mech., 27, pp. 541– Elsevier, Amsterdam, 2004.
549.
[18] Piomelli, U., Balaras, E., and Pascarelli, A., 2000, [26] 44. Sagol, E.; Reggio, M.; Ilinca, A. Issues concerning
“Turbulent Structures in Accelerating Boundary Layers,” J. roughness on wind turbine blades. Renew. Sustain. Energy
Turbul., 1, pp. 1–16. Rev. 2013, 23, 514–525.

[19] Lanzafame, R.; Mauro, S.; Messina, M.Wind turbine [27] Villalpando, F.; Reggio, M.; Ilinca, A. Numerical study
CFD modeling using a correlation-based transitional of flow around iced wind turbine airfoil. Eng. Appl.
model. Renew. Energy 2013, 52, 31–39. Comput. Fluid Mech. 2012, 6, 39–45.

[20] Fatimah M. Mohsen, Hameed k. Hamzah, Dhragham [29] Thangam, S., & Speziale, C. G. (1992). Turbulent flow
A.Alkhafaji (2021). CFD Based Investigation of Boundary past a backward-facing step-A critical evaluation of two-
Layer Separation Control for Y-intake Duct Adapter by equation models. AIAA journal, 30(5), 1314-1320.
Various Degrees of Height Roughness. Journal of
Mechanical Engineering Research and Developments ISSN:
1024-1752 CODEN: JERDFO Vol. 44, No. 2, pp. 371-382
Published Year 2021

[21] Jessam, R. A., Al-Kayiem, H. H., & Nasif, M. S.

(2018). Experimental and numerical analysis of different
flow modifier on the reversal flow region in s-shaped
aggressive diffuser. AIP Conference Proceedings, 2035, 1–
8.

[22] Yakhot, V., Orszag, S.A., Thangam, S., Gatski, T.B. &
Speziale, C.G. (1992), "Development of turbulence models
for shear flows by a double expansion technique", Physics
of Fluids A, Vol. 4, No. 7, pp1510-1520.

[23] Martin Länsmans, Anders Ramqvist. Computational

fluid dynamics study of the spillway and plunge pool at
Baihetan hydropower station. KTH School of Industrial
Engineering and Management Energy Technology
EGI_2017-0053 EKV1194 Division of Heat and Power
Technology SE-100 44 STOCKHOLM.