MECH 6480: COMPUTATIONAL FLUID DYNAMICS CFD
ASSIGNMENT 1 - Simulation of incompressible flow
1 PREAMBLE
The aim of this activity for you to gain experience in conducting a holistic analysis of an incom-
pressible flow problem. To achieve this you will perform the following tasks and then submit
and engineering report detailing your work together with a folder containing the corresponding
simulations. With this assignment we are trying to expose you to the following activities:
1. creating a mesh that resembles a wing being tested in windtunnel test;
2. verify your simulation approach against windtunnel test data;
3. conduct a gridding independence study and explored effects of using different turbulence
models;
4. use CFD to vary an aspect of the wing and see how this affects the wing performance;
and
5. use CFD to explore how the same wing (original or modified) performs in free flight.
This is an individual assignment, and you will have to get certain individualised simulation
parameters from Blackboard.
To assure a successful outcome of your assignment you are advised to complete steps 1 and 2
as soon as possible and we (tutors and teaching staff) will be able to provide personal feedback
during the tutorials.
2 INTRODUCTION
For this project is centred around the Clark-14 aerofoil, which many of you will have tested as
part of MECH3410 (or at least attended the virtual practical). The aerofoil installed in the
wind-tunnel, together with critical dimensions is shown in Fig. 1. To complete the assignment
you will need to some experimental data. This is available on Blackboard in Experimental
Data.pdf. In addition, to keep the assignment interesting, we are varying some parameters
between students. Please have look at the table on Blackboard to see which Case and Variation
has been assigned to you. Here Case refers to one of the test conditions from the experimental
data set. You should use the corresponding Angle of Attack (AoA) and Velocity for all your
simulations. And Variation refers to what should be adjusted during step 5 of the assignment.
The remainder of this document is structured to give you some additional guidance on
each of the five tasks, before finishing with a checklist on what to include in the report. The
corresponding assessment rubric is available on Blackboard.
3 TASKS
3.1 1: create a mesh
Create a CFD grid for incompressible viscous flow analysis of the wing in the windtunnel. This
mesh needs to include the wing and the windtunnel walls. The extend of the domain should
match Fig. 1.
1
Figure 1: Wing installed in the windtunnel and critical dimensions
The aerofoil is a Clark Y-14. To get an aerofoil profile you can use an number of online
tools, e.g. http://airfoiltools.com/plotter/index, but make sure you scale the thickness
so that you obtain a maximum thickness of 14 of chord length.
This grid should be smoothed and have appropriate refinement in the areas where pressure
and velocity gradients are high. This includes refinement along the walls, especially in regions
where you want to resolve the boundary layer and appropriate y+ value should be achieved.
Keep in mind that grid refinement is an iterative approach. Hence you will have to start
running some simulations and then return to refine your mesh again.
3.2 2: verify simulation approach against experimental data
Step 2 and step 3 are somewhat intertwined, hence expect to work on these tasks concurrently.
There is a specific difference between this and the next task. In task 2 the aim is to show
that your results are consistent with the experimental data. For the current experiment the
data you have is pressure at discrete points along the wing surface. Hence you want to show
that the predicted pressures match the wing pressure distribution and to use the data to identify
possible shortcomings in your simulation approach. This means you are not looking for an exact
match at each location. You will actually discover that simulations and experiments rarely
match. Instead you are looking for a qualitative agreement and possible regions significant
disagreement. Identifying regions where differences occur allows you to identify the limitations
of your modelling approach.
For this step we expect you extract pressure along the wing surface and to plot this against
chord location. This allows the pressure data from the experiment to be overlaid on top of
these traces. Ideally you can show that the two data match. You may see discrepancies on the
top, especially if the flow separates and this is something to explore and consider when varying
turbulence models.
3.3 3: grid independence and choice of turbulence model
Whereas the previous step was about showing whether the model correctly captures the flow
physics (i.e. show that you are using the correct model), this step is about showing that your
simulation approach doesn’t introduce undue numerical errors.
2
One dominant cause of the numerical errors is discretisation. This can be assessed by con-
ducting a grid independence study. For this you conduct simulations on three grids, ideally
halving grid spacing every time. Using the parameters of interest (Lift coefficient and Drag co-
efficient for a wing) then allows these to be tabulated or plotted to show that these parameters
are converging to a value. For extra reading and how to calculate the Grid Convergence Index
(GCI), have a look at https://fluiddynamics.uqcloud.net > MECH6480 > OpenFOAM
> Examining Convergence (e.g. grid dependency study) - https://fluiddynamics.uqcloud.
net/#/resource/5a0ba18d34cc363763e05e99/5d3fa1a0693d0c4895fd5ade/5f43b4718eff639a1b9bb20e/
5f65ecb18eff639a1b9bb2dd.
For this step you should conduct simulations using at least three progressively refined meshes
and then analyse the predicted values of CL and CD to establish what mesh size is suitable for
further simulations.
Next you should also investigate different modelling approaches for momentum closure.
Three possible options are laminar, turbulent - Spallart Almaras, and turbulent - k-epsilon. Us-
ing the grid-converged mesh explore at least two modelling approaches and then by considering
the experimental data from step 2 make an appropriate choice of which turbulence model gives
the best results.
3.4 4: exploring geometric variations
For this step you are asked to vary the geometry to the wing. Please apply the change that has
been specified as part of your instructions from Blackboard.
For thickness variations, simply scale the wing thickness. For chamber line variations,
simply apply curvature to the wig profile.
In this step you make use of CFD’s ability to easily adjust model geometry and to explore
how this influences wing performance without the need to make new physical models. Especially
if you have defined you geometry and mesh parametrically, it can be very easy to explore a
wide range of the design space.
3.5 5: free flight performance estimate
Finally, you want to know how the wing performs in free-flight. That is how does the lift
generation and drag change, once the constraints of the wind-tunnel walls, which constrain the
flow direction above and below the wing are removed. This is easily done by changing the
boundary conditions away from wall to appropriate free-stream conditions.
For this step re-simulate your geometry from step 3 or 4 with a grid independent mesh and
using free-stream boundary conditions. Explore how lift and drag and the flow field change.
4 REPORT CHECKLIST & SUBMISSION
4.1 Report
The overall report should start with an executive summary, followed by an introduction that
set out the aims of the study you report. Finally there should be a conclusion highlights what
you have learned about the applying CFD in the design process of aerodynamic components.
This should highlight advantages and possible pit-falls. The report should be no more 20 pages.
The main section should follow the five steps outlined above. As a minimum this should cover
the following:
Brief description of numerical method being used. This should also explain why the
method(s) are appropriate.
Details on why you have chosen the extend of the fluid domain being modelled, what
boundary conditions you have selected, and why these are appropriate. Also include a
description of the blocking structure and mesh refinement that has been applied and use
grid quality metrics to confirm that your mesh is good.
3
Discuss the results from your verification exercise. For this discuss how well CFD and
experiment agreed and what discrepancies were observed. For the discrepancies that exist,
provide some discussion about their possible causes.
Describe the details of your grid dependency study and use the results to identify a mesh
with a sufficient number of cells for the remainder of the analysis.
Discuss the turbulence model that has been selected. This should include any specific
turbulence model settings you have selected (including initial and boundary conditions).
Describe how you have varied the wing geometry and how this has altered performance
and the flow field.
Discuss how the fluid domain and boundary conditions were altered for the free-flight
case. Then discuss the changing these simulation parameters affected your results.
For the comparison between windtunnel and free-flight results you should look a both
at the pressure distribution over the wing and how the flow-field further away from the
wing is affected. Taking vertical slices through the fluid domain some distance up and
downstream of the wing (and maybe some horizontal slices too) and assessing how pressure
or flow direction has been altered may be a good way to do this.
Please keep in mind that this report is not just a box ticking exercise. The above is a list of
minimum content. When marking the report and the submitted files (see below) we will look
at this as a whole package of work. To achieve high marks you need to show that you are
competent in applying CFD to problems like the one specified, that you have the skills to apply
each of the steps in a competent manner, and that you can communicate the findings from each
step at the required standard.
The report will be submitted via Blackboard.
4.2 File Submission
In addition must create the following on your VM. (For students that are not using the course
provided VMs, you will have to log into your course VM and copy across the files.
Create a directory called Assign-1 on the desktop of your VM. We will download a copy
of this directory so that we can assess and mark your simulation work directly. In this
directory create copies of the following:
Three simulation Cases corresponding to the three grid refined simulations from step 3.
These should include the simulation results.
Simulation Case from Step 4. This should include the simulation results.
Simulation Case from Step 5. This should include the simulation results.
Make sure all five case directories have been labelled clearly and all the files to run each of
these cases are included. We will re-run or review the results from some of your cases, so make
sure these cases run.