Journal of Physics: Conference Series

PAPER • OPEN ACCESS

Effect of applying screen and honeycomb to the flow characteristic in

wind tunnel based on CFD simulation
To cite this article: P A Fadilah and D F Erawan 2018 J. Phys.: Conf. Ser. 1130 012008

View the article online for updates and enhancements.

This content was downloaded from IP address 139.81.91.177 on 29/11/2018 at 15:07

6th International Seminar of Aerospace Science and Technology IOP Publishing
IOP Conf. Series: Journal of Physics: Conf. Series 1130 (2018)
1234567890 ‘’“” 012008 doi:10.1088/1742-6596/1130/1/012008

Effect of applying screen and honeycomb to the flow

characteristic in wind tunnel based on CFD simulation

P A Fadilah , D F Erawan

Chroma International

Abstract. In the field of wind tunnel testing, flow quality is one of considerations which
determine the accuracy of data measurement. There are several aspects that affect the flow
quality, such as blade design, flow conditioner application and tunnel circuit configuration (i.e.
design of the corner vane, corner duct, contraction duct and etc.). Applying flow conditioner
like screen or mesh and honeycomb is commonly used to improve the flow quality in the
tunnel. This paper discusses about the effect of applying screen and honeycomb to the flow
characteristic in a simple tunnel model. The screen was applied on the several locations at the
WAD (wide angle diffuser), i.e. at the inlet, middle section, and at the outlet of the WAD.
Honeycomb was applied after the last screen in the tunnel duct. In the modelization, the applied
screen or mesh and honeycomb were modeled using porous media computational model and
using Darcy law as a mathematical model. The CFD simulation was performed using Numeca
by separating the domain into fluid domain section (for the tunnel section) and porous media
domain (for the screen and honeycomb section). Unstructured mesh type, generated by
Hexpress, was used on the tunnel computational domain. The results indicate the reduction of
flow separation, hence more uniform flow in the tunnel, by applying screen or mesh at the
WAD and by applying honeycomb in the chamber.

1. Introduction
Screen or mesh is commonly used in the field of wind tunnel to solve certain flow problems such as
the turbulence intensity and flow recirculation. Wind tunnel with WAD (Wide Angle Diffuser) has
usually problem with the flow quality due to the disturbance that occurs when the flow goes through to
the WAD and run into separation. To overcome this problem there are several methods which can be
applied, for example by applying splitter in the WAD section or applying screen or mesh of which one
of the drawbacks is the increase in pressure loss.
Referring to previous studies of perforated plate, screen or mesh it has been shown, for example,
according to Charles L. Wharton, screen or mesh can be applied in the wind tunnel to prevent flow
separation at WAD [1]. It is also common to use screen together with honeycomb as flow conditioner
according to Louis Cattafesta [2]. Tensioned screens are placed in the settling chamber to break up the
larger-scale turbulent eddies into a number of small eddies. The purpose of this study is to better
predict the screen or mesh and honeycomb effect on the flow characteristic at the wind tunnel wide
angle diffuser (WAD). This kind of prediction is helpful in wind tunnel manufacturing when a WAD
is needed in a circuit. In this study the computational model consists of 3 components, i.e. propeller,
diffuser (including WAD) and a chamber. Using this kind of model, we can study the effect of honey

Content from this work may be used under the terms of the Creative Commons Attribution 3.0 licence. Any further distribution
of this work must maintain attribution to the author(s) and the title of the work, journal citation and DOI.
Published under licence by IOP Publishing Ltd 1
6th International Seminar of Aerospace Science and Technology IOP Publishing
IOP Conf. Series: Journal of Physics: Conf. Series 1130 (2018)
1234567890 ‘’“” 012008 doi:10.1088/1742-6596/1130/1/012008

comb and mesh to the angularity (due to swirl behind propeller) and flow separation (at the WAD).
For modelling the screen in the computational simulation, porous media model was used. This model
is provided in Numeca based on Numeca user guide [5] of which the pressure drop can be calculated
using different laws alternatively, such as Darcy law, Ergun Law, Integral Law. In the porous media
relation there are several variables that influence the pressure drop and resistance of the
grid/screen/mesh. Idelcik concluded that the resistance of a perforated plate (similar with
grid/mesh/screen) is correlated to the plate/grid porosity, Reynolds number and plate/grid thickness (t)
[3]. Another study that was conducted by Katarina Nilsson discussed about the effect on porosity and
plate thickness wich can affect the pressure drop [4].

2. Method of analysis
The method of analysis which is used in this study consists of analytic method and numerical method.
The analytical method is discussed later. It contains several equations that explain the relation between
velocity, pressure, pressure loss, etc. The numerical method use Numeca software as CFD
(Computational Fluid Dynamic) to simulate the flow and to predict the effect of screen/mesh or
honeycomb application. In this study there are several cases that were simulated and analyzed as
summarized in the following table below.

Table 1. Studied cases.

Screen
Cases Honeycomb
Typea Number
1 w/o w/o w/o
2 A 3 w/o
3 B 3 w/o
4 C 3 w/o
5 D 3 w/o
6 C 1 w/o
7 C 2 w/o
8 C 3 1

a
Screen model A,B,C,and D is discussed in the table below.

Table 2 summarizes the screen/mesh and honeycomb type which were used in this study, consisting of
four differrent screen porosity between 71% and 47 %.

Table 2. Screen/mesh type.

Screen/Mesh Type Porosity (%)

Screen A 71
Screen B 63
Screen C 57
Screen D 47
Honeycomb 80

2.1. Porous media

To make a computation model of the screen and honeycomb, a porous media model is generated. The
model can representating the actual orifice at screen and honeycomb that difficult to be meshed in
actual geometry model. As defined in NUMECA documentation that a porous medium is a solid
permeated by an interconnected network of pores filled with a fluid. Usually both the solid matrix and
the pore network are assumed to be continuous, so as to form two interpenetrating continua such as in

2
6th International Seminar of Aerospace Science and Technology IOP Publishing
IOP Conf. Series: Journal of Physics: Conf. Series 1130 (2018)
1234567890 ‘’“” 012008 doi:10.1088/1742-6596/1130/1/012008

a sponge. Many natural substances such as rocks and soils, and man-made materials such as foams and
ceramics can be considered as porous media. When simulating the flow in porous medium the user has
to use a porous media model because the geometry is too complex to resolve with a grid [10].

2.2. Analytical method to modelize screen/mesh by porous media

The analytical method was done based on the pressure equation and relation as shown in the equation
(1) to (6). Meanwhile the pressure drop is calculated based on screen specification such as porosity,
wire diameter—which is assumed to be similar to the thickness of the screen/mesh—and diameter of
the orifice; also velocity through the screen. The equation to calculate pressure drop through the screen
is derived as follows [3],

(1)

(2)
⁄

Kf is function of Re and . For high Re (Re > 200,000) Kf is equal to zero . Therefore, for high Re,
the equation becomes

(3)
The value of the constant in the high Re is equal to 0.98

(4)
( )
⁄

where is friction factor which for high Re it can be assumed to be close to zero, therefore equation
(4) becomes

⁄ (5)

) ) ) (6)

where  is function of t/d as shown in [3] and  is porosity.

As previously mentioned, this study analysed the effect of applying screen in the WAD at the three
differences locations, i.e. at inlet, at middle section and at the outlet of the WAD. Based on equation
(1), pressure drop is function of velocity therefore the pressure drop for each section WAD are
differents.

2.3. Numerical method

The numerical method, as previously discussed, is based on CFD (Computational Fluid Dynamics)
with steps as the ones in the standard CFD procedure. It started with making computational geometry
model, then mesh generation, followed by solving the Navier-Stokes equations and then post
processing. In the process of building computational geometry model using CAD software (for
example CATIA or Solid Works), the tunnel or duct model was separated into several sections,
consisting of fluid block section, porous media section and rotating domain section. Fluid block
section is the modeling of the fluid flow contained in the tunnel/duct with certain boundary conditions,
the porous media section represents the screen/mesh by applying the Darcy law based on pressure loss

3
6th International Seminar of Aerospace Science and Technology IOP Publishing
IOP Conf. Series: Journal of Physics: Conf. Series 1130 (2018)
1234567890 ‘’“” 012008 doi:10.1088/1742-6596/1130/1/012008

equation, the last section is rotating domain which modeled the rotation of the model blade.

3. Simulation modeling
3.1. Computational geometry model
The geometry of the duct/tunnel with several parts of domain are depicted in the figure 1 below. As
described previously, the computational model contains fluid block domain, porous media block
domain which represent the screen/mesh and honeycomb; and rotating domain which models the
rotation from the fan blade.

Screen 1
Screen 2
Screen 3 Honeycomb

Figure 1. Computational model of the duct/tunnel.

3.2. Meshing
The meshing that was applied on the computational model uses structured mesh type and unstructured
mesh type (cartesian mesh). The structured mesh type was applied in the rotating domain and the
unstructured mesh was applied on the stationary parts, on the tunnel, screen, and the honeycomb. The
rotating domain was meshed with Autogrid-5, a turbomachinery platform and a structured mesh
generator with good quality of mesh result. The mesh that was used in the study is depicted in figure 3
below.

Figure 2. Applied mesh on the rotating domain.

4
6th International Seminar of Aerospace Science and Technology IOP Publishing
IOP Conf. Series: Journal of Physics: Conf. Series 1130 (2018)
1234567890 ‘’“” 012008 doi:10.1088/1742-6596/1130/1/012008

Figure 3. Applied mesh on the duct/tunnel computational domain.

3.3. Boundary conditions

The computational simulation started with the generation of computational domain in which the flow
properties such as velocity, pressure and temperature would be calculated. The domain was defined
with some boundaries such as inlet, outlet, connecting domain and porous media domain as shown in
figure 4. The boundary conditions that were applied in the pre-setting of the simulation steps were
condition on the inlet, wall and outlet. For the inlet we used mass flow and input RPM to define the
rotational speed of the model blade; for the wall we used default setting of the wall boundary
conditions and for the outlet we used static pressure as input value. The Darcy law was applied in the
porous media domain setting. As previously stated, in this simulation the screen are modelled as
porous media. To set input for porous media options in the NUMECA software, we must define the
Value K’, where for constant value of K’, K’ is defined as the following equation (defined as in the
NUMECA user guide [5]):
⃑ (7)

To accomodate the screen thickness, the equation to determine value of K’ is modified to be

(8)

where t is the thickness of the screen model, and v is the flow velocity that is prependicular to the
screen area. The location of the boundary conditions that were applied in the computational domain is
depicted in figure 4 below.

5
6th International Seminar of Aerospace Science and Technology IOP Publishing
IOP Conf. Series: Journal of Physics: Conf. Series 1130 (2018)
1234567890 ‘’“” 012008 doi:10.1088/1742-6596/1130/1/012008

Figure 4. Boundary that was applied on the computational geometry model.

3.4. Numerical model

The equation that was used in the simulation is turbulent Navier-Stokes with K-epsilon (extended wall
function) turbulence model. The number of mesh that was applied on the computational domain
geometry was automatically defined by Numeca software. The type of the simulation is steady
simulation and optional model of porous media was chosen. The porous media was used to simulate
the screen/mesh. It was then applied at the domain where the screen/mesh was located.

3.5. Convergence of the simulation

The criteria of convergence which were used in the simulation were residual growth of mass and
momentum as shown in figure 5. The residual is one of the most fundamental measures of an iterative
solution of convergence, as it directly quantifies the error in the solution of the system equation. In a
CFD analysis, the residual measures the local imbalance of a conserved variable in each control
volume. For complicated problems, however, it is not always possible to achieve residual levels as low
as -6 (global residual scale in NUMECA). However, if the global residual curve cannot reach a certain
low level, the convergence of the simulation still can be assessed by monitoring the integrated
quantities, such as force, drag, mass flow or average temperature. The simulation is convergence when
the integrated quantities value does not fluctuate much anymore. Figure 3 shows the convergence
history of integrated quantities that is represented by mass flow.

Figure 5. Mass flow confergence hystory.

6
6th International Seminar of Aerospace Science and Technology IOP Publishing
IOP Conf. Series: Journal of Physics: Conf. Series 1130 (2018)
1234567890 ‘’“” 012008 doi:10.1088/1742-6596/1130/1/012008

4. Simulation result and discussion

This section discusses the results of the simulation of each case, starts from case 1 to case 8. Each case
is explained briefly with several visualisation like velocity vector and several contours.

4.1. Simulation without screen/mesh and honeycomb applied with fan blade on
This section discusses about the flow characteristic when the fluid flows through to the tunnel/duct
without screen/mesh treatment. The flow characteristic of the flow in the tunnel/duct is depicted in
figure 6 below.

Figure 6. Streamline/flowpath in the tunnel without any screen and honeycomb applied.

Figure 6 shows that the flow runs into massive separation when it passes the WAD. This could be
explained by the fact that the diffuser opening is too wide that it provokes separation. The
recirculation flow caused by the separation is shown by the irregular flow paths/streamline.

4.2. The effect of the screen/mesh porosity

In table 3 below the comparison from each screen to the ability in reducing flow separation at the
WAD section is shown.

Table 3. Comparison of streamline/flow path of each tunnel/duct configuration.

Screen Type Simulation Result

A (Porosity 71%)

B (Porosity 63%)

C (Porosity 57%)

7
6th International Seminar of Aerospace Science and Technology IOP Publishing
IOP Conf. Series: Journal of Physics: Conf. Series 1130 (2018)
1234567890 ‘’“” 012008 doi:10.1088/1742-6596/1130/1/012008

D (Porosity 47%)

Based on the simulation result that are shown in the table 3, it can be deducted that increasing
screen/mesh porosity reduces the flow separation at the WAD (Wide Angle Diffuser) section. There is
siginificant effect when the screen porosity is changed from porosity of 71% to porosity of 63%,
massive separation that occur at the WAD section before, significantly thinned off by changing the
screen/mesh porosity. This pheomena indicates that applying screen/mesh is useful to reduce the flow
separation.

4.3. The effect of the number of screen/mesh

In this section the effect of number of screens to the ability in reducing the flow separation phenomena
at the WAD section is assessed. Screen/mesh which is used in this section is screen/mesh of type C
(with 57% porosity). Table 4 here below shows that separation and recirculation flow are reduced
progressively by increasing screen/mesh number. Initially, with one screen/mesh the flow separation is
still massive. Then by adding another screen/mesh which was applied in the duct, the separation
become less massive. With three screen applied in tunnel/duct the separation seems to completly
disappear especially after screen number 3.
Table 4. Comparison of streamline/flowpath of each configuration by varying the
screen/mesh number.

Screen Type Simulation Result

1 Screen Applied

2 Screen Applied

8
6th International Seminar of Aerospace Science and Technology IOP Publishing
IOP Conf. Series: Journal of Physics: Conf. Series 1130 (2018)
1234567890 ‘’“” 012008 doi:10.1088/1742-6596/1130/1/012008

3 Screen Applied

As previously discussed, based on the simulation result that is presented in the table 4, the effect of the
addition of screen’s number has similiar effect with increasing screen/mesh density (lowering the
porosity number). At the one screen mesh activated case, there is still occur massive recirculating flow
at the WAD. With increasing active number of screen/mesh, the recirculating flow become thinned off
until it almost disappear when the three screens are all activated.

4.4. The effect of applying honeycomb

In this section the effect of adding honeycomb in the tunnel/duct to the flow characteristic is discussed.
The table 5 and table 6 here below depict the flow using the streamlines in the first figures and
velocity vector distribution in the second. From the simulation result in table 5 it seems there is no
significant difference of the stream line/flow path in the tunnel from the two configurations.

Table 5. Comparison of streamline/flow path from configuration with and without honeycomb.

Screen Type Simulation Result

Without
Honeycomb

With Honeycomb

However from the figures in table 6 below, which show the distribution of velocity vector at the

9
6th International Seminar of Aerospace Science and Technology IOP Publishing
IOP Conf. Series: Journal of Physics: Conf. Series 1130 (2018)
1234567890 ‘’“” 012008 doi:10.1088/1742-6596/1130/1/012008

downstream section after honeycomb, we can notice the difference of the velocity vector distribution.
The tunnel/duct with honeycomb has more uniform velocity vector distribution than the one without it.
Table 6. Comparison of velocity vector distribution of tunnel/duct with and without
honeycomb configuration.

Screen Type Simulation Result

Without
Honeycomb

With Honeycomb

5. Conclusions
In this study the effect of screen/mesh and honeycomb that were applied in the tunnel/duct with fan on
condition was assessed. First, it can be concluded that the screen/mesh can prevent the flow separation
at the WAD section. The ability to reduce flow separation increases with the porosity of the screen,
though this increase of porosity causes the pressure loss due to the increase of the number of the
screen itself. In the study of honeycomb effect it is shown that by applying honeycomb in the
tunnel/duct configuration we can obtain more uniform flow. The compromise between separation area
and pressure loss must be studied and the best solution will depend on the configuration of the tunnel.
For this example of configuration, it is found that the best one was using three screeens with 57%
porosity (screen type C) at the WAD section and applying honeycomb after the WAD’s last screen.

References

[1] Charles Lancaster Wharton, The Effect of Screens in a Wide Diffuser of Square Cross-Section,
Faculty of the Graduate Division Georgia Institute of Technology, 1954.
[2] Louis Cattafesta, Fundamentals of Wind Tunnel Design, Department of Mechanical and
Aerospace Engineering, University of Florida , Gainesville, FI, USA, 2010.
[3] L.J. Weber, M.P. Cherian M.E. Allen, and M. Muste, Head loss characteristics for Perforated
Plated and Flat Screens, IIHR Techinical Report No. 411, Lowa Institute of Hydraulic
Engineering.
[4] Katarina Nilsson, Departmen of Chemical Engineering, Study of Pressure Drop for Perforated
Plates and TAD-Fabrics, Lund Institute of Technology.
[5] James Scheiman, Consideration for the Installation of Honeycomb and Screen to Reduce
Wind-Tunnel Turbulence.NASA Techincal Memorandum 81868, August 1981.
[6] A. Szymkiewics, Modelling Water Flow in Unsaturated Porous Media, GeoPlanet : Earth and
Planetary Sciences, DOI : 10.1007/978-3-642-23559-7_2, Springer-Verlag Berlin
Heidelberg 2013.

10
6th International Seminar of Aerospace Science and Technology IOP Publishing
IOP Conf. Series: Journal of Physics: Conf. Series 1130 (2018)
1234567890 ‘’“” 012008 doi:10.1088/1742-6596/1130/1/012008

[7] Channing R.C., Ted Chu, and K.H. Wang, Study of the Head Loss Associated with a Fluid
Flowing through a Porous Screen. Department of Civil and Environmental Engineering,
University of Houston, texas
[8] Emrah Ozahi, An Analysis on the pressure loss through perforated plates at moderated Reynold
Number in Turbulent Flow Regime, University of Gaziantep, Faculty of Engineering,
Mechanical Engineering Department, 27310. 2015 Elsevier.
[9] Flow Through Porous Media,www.compassis.com
[10] NUMECA user guide/documentations/theory guide/porous Media Equation
[11] https://en.wikipedia.org/wiki/Flow_separation

11