School of Mechanical and Manufacturing Engineering

Assignment Submission
Student Name(s): Shivam Goel

Student Number(s): 19212660

Programme: M.Eng in Mechanical & Manufacturing Eng

Assignment Title: Assignment Part A Impulsively Flat plate

Module code: MM432

Lecturer: Dr Yan Delauré

Project Due Date: 30/10/2019

Declaration

I declare that this material, which I now submit for assessment, is entirely my own work and has not been
taken from the work of others, save and to the extent that such work has been cited and acknowledged within
the text of my work. I understand that plagiarism, collusion, and copying is a grave and serious offence in the
university and accept the penalties that would be imposed should I engage in plagiarism, collusion, or copying.
I have read and understood the Assignment Regulations set out in the module documentation. I have identified
and included the source of all facts, ideas, opinions, viewpoints of others in the assignment references. Direct
quotations from books, journal articles, internet sources, module text, or any other source whatsoever are
acknowledged and the source cited are identified in the assignment references.

I have not copied or paraphrased an extract of any length from any source without identifying the source and
using quotation marks as appropriate. Any images, audio recordings, video or other materials have likewise
been originated and produced by me or are fully acknowledged and identified.

This assignment, or any part of it, has not been previously submitted by me or any other person for assessment
on this or any other course of study. I have read and understood the referencing guidelines found
at http://www.library.dcu.ie/citing&refguide08.pdf and/or recommended in the assignment guidelines.

I understand that I may be required to discuss with the module lecturer/s the contents of this submission.

I/me/my incorporates we/us/our in the case of group work, which is signed by all of us.

Signed: _________________________
Part 1:
Answer the following questions to assist with the model set-up. All answers should
be justified:
a. Why is the mesh cell height reduced gradually with the distance to the bottom boundary?
Note that in CFD, the governing equations are solved according to the Control Volume Analysis
method, each cell representing a Control Volume (CV).
Cell height:
The mesh cell height reduced gradually with the distance to the bottom boundary because most of the
variations in velocity are happening near the bottom boundary and each cell represents a control
volume (CV), for the accurate results we need smaller control volumes.
b. Is the large aspect ratio of the cell next to the bottom wall an issue? You should again refer to
the fact that each cell is a Control Volume and that the equations are solved over this CV
according to the Control Volume Analysis methodology. You can also search the Fluent user
guide for guidelines on maximum recommended aspect ratio.
Aspect ratio:
Aspect ratio is a measure of the stretching of the cell. The quadrilateral/hexahedral/wedge cells inside
the boundary layer, on the other hand, can be stretched to aspect ratio of up to 10:1 in most cases.
With regard to the stability of the flow solution, it can go as high as possible; however, with regard to
the stability of the energy solution, the maximum aspect ratio should be kept below 35:1. [1]
Large aspect ratio is not an issue because the mesh is quadrilateral and since the mesh cell height is
changing towards the bottom boundary.
c. What type of viscous model (Laminar or Turbulent) should be used and would this allow
comparison with known analytical solutions?
Viscous Model:
As stated in the question, laminar model is used for CFD.
Yes, this will allow comparison with known analytical solution. For analytical solution Reynolds
number can be used to check whether the flow is laminar or turbulent. The velocity of the fluid is 1
m/s with dynamic viscosity of 1.4607e-5 which gives Reynolds number of the range confirming it as
laminar.
d. An unsteady solver should be used, explain why. Also explain how the time step can be
chosen? To do this, think about what is assumed about fluxes at the CV boundaries in the
derivation of the governing equations. Note that if the time step is too large your solution will be
unstable or inaccurate. So, to help you answer this question, think about what it means in CFD
when the time step is increased.
Unsteady solver:
Transient option has chosen to solve this problem as the problem is unsteady, it depends on time.
Time step is the number of steps which has been chosen to know what’s happening to the flow and get
a better picture. If the time step is too large, we might miss the important relevant information. So, we
need to perform iterations to get perfect time steps for accurate results.

1
e. State which boundary conditions you chose for the top and bottom boundaries and explain
why. To do this you should review the options available in Fluent and justify your choice from a
purely logical perspective. For the left and right boundary, you should impose periodic
boundary conditions as described in the “Periodic Heat Flow” Fluent Tutorial, explain why?
Boundary conditions:
Top boundary layer has been made stationary while bottom boundary layer movable at a velocity of 1
m/s. This is because it’s given in the question that the plate is moving in the horizontal direction.
Left boundary has been made periodic boundary and right as shadow zone because the fluid is
entering from the left boundary and leaving from the right.

f. State which values you used to initialize the velocity and pressure fields and
discus why.

Velocity and pressure fields:

Initial velocity has been kept at zero because plate started impulsively before that it was at rest.
Pressure is also taken as zero due to the assumption of standard atmospheric pressure which makes
gauge pressure as zero.
g. Explain why gravity effects can be neglected if you are only interested in
finding the velocity distribution in the flow field (that is if you want to know
how velocity changes with position in space).

Gravity effects:
Gravity doesn’t affect in determining the velocity distribution of the flow field as the fluid is flowing
horizontally over the plate and gravitational force acts in the vertical direction. Also, the gravitational
force is evenly distributed on the plate so it will not affect the flow so it can be neglected.
h. Explain if you can assume that the fluid density is constant. (assume that
you are dealing with standard atmospheric conditions)? You can use a
segregated pressure-based solver to model this problem.
Fluid density:
Air is the fluid here. It can be compressible or incompressible depending upon the flow of speed. If it
is flowing at lower speeds (less than 100 m/s) [2] it is considered as incompressible so it can be
assumed that fluid density is constant.

2
Part 2:
Once the problem is setup run the solution and extract the velocity profile in the
boundary layer above the plate. You should use the SIMPLEC solver and the
default Residual thresholds (10−3 for all residuals) for this question. Compare the
computational results with a known analytical solution at 1.5s from the initial time.
To help with the comparison, you should provide XY-plots of the normalized
solution. The computational and exact analytical solutions should be plotted on a
single graph. This can be done in Excel. The analytical solution can be found in
most text book on Fluid Mechanics (e.g. P.K. Kundu and I.M. Cohen, 2004, Fluid
Mechanics, Pub. Elsevier Academic Press) and is given by:

𝒖 𝒚
= 𝟏 − 𝒆𝒓𝒇 [𝟐 ]
𝑼 √𝝂𝒕

Answer:

Velocity Profile ANSYS vs Analytical

2.5

1.5
ETA

0.5

0
0 0.5 1 1.5
u/U

analytical ansys

Above graph shows comparison between computation results using ANSYS and analytical method.
XY graph has been plotted with ETA on y-axis, u/U on x-axis, orange line represents results from
ANSYS and blue line represents analytical results.
There is visible difference between the two graphs as ANSYS curve converges more in comparison to
analytical curve. To make ANSYS solution more accurate we should make some changes in model
set-up.

Part 3:

3
Use the above data to discuss whether the computational results are accurate.
Propose improvements to the model set-up if needed and justify your choices. You
should consider:
a. The order of accuracy of the discretisation schemes.
b. The residual Thresholds.
c. The time step
You should search the literature (including the Fluent User/Theory Guide) for an
explanation of the meaning and significance of these parameters. Test whether the
proposed changes improve the results and provide an assessment.
a. The order of accuracy of the discretisation schemes.

Order of Accuracy
3

2.5

2
ETA

1.5

0.5

0
0 0.2 0.4 0.6 0.8 1 1.2
u/U

analytical ansys second order

In the above figure, three curves have been plotted. Orange curve is made using analytical method and
yellow has been made using first order upwind in ANSYS while brown curve with second order
upwind in ANSYS.
To compare with analytical solution, initially first order upwind method has been used. As we can see
the curve converges more from analytical solution, second order upwind method is used for better
accuracy. The curve made using second order upwind method has a slight difference as compared to
first order upwind method but more accurate.

4
b. The residual Thresholds

Residual threshold
3

2.5

2
ETA

1.5

0.5

0
0 0.2 0.4 0.6 0.8 1 1.2
u/U

analytical ansys 0.0001

In this, residual values have been varied to get more accurate results. It has been demonstrated in the
above graph that lesser residual values give more accurate results. Residual value of 0.0001 is closer
to the analytical solution as compared to residual value of 0.001.

5
c. The time step

Time step
3

2.5

2
ETA

1.5

0.5

0
0 0.2 0.4 0.6 0.8 1 1.2
u/U

analytical ansys time step 0.01 time step 0.0075

In this, different time steps have been chosen to get accurate results.
It has been illustrated here that large time step that is 0.01 and small time step that is 0.005 have given
less accurate results. After performing trial and error, time step of 0.0075 has given results closer to
analytical results.

6
References:
[1] ANSYS Fluent Guide, [online],
http://www.afs.enea.it/project/neptunius/docs/fluent/html/ug/node167.htm

[2] Çengel Y A., Cimbala J M., Turner R H., Fundamentals of Thermal Fluid Sciences, (5th ed.),
McGraw Hill Education, New York, p. 420