Menu

Discussion Forum

New Discussion
Filter

Discussion Closed This discussion was created more than 6 months ago and has been closed. To start a new discussion with a link back to this one, click here.

Time Dependent Study of an Airfoil- Lift Coefficient

Posted Apr 2, 2013, 1:19 a.m. EDT Fluid & Heat 5 Replies

Ting Lau
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Hi,

I am trying to learn how to calculation the lift and drag coefficient of an airfoil running a time-dependent study.

Below are the boundary conditions.

Re ~2000
Inlet- Inlet U*step1(t[1/s])
Top and bottom- Slip
Outlet- Pressure no viscous stress

My problems the Lift coefficient (CL) values seems like it keep going down as time goes.
Re~2000 but would like to increase it to 10^7.

I am not sure what's wrong with my setup?
I set it up based on the model in the Library (Flow past a cylinder). File is too large cannot be attached so screenshots are attached instead.

Thank you very much for your time!

Ting
5 Replies Last Post Nov 22, 2016, 3:38 a.m. EST
Josh Thomas Certified Consultant
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago Apr 2, 2013, 9:23 a.m. EDT
Ting-

For flows with these Reynolds numbers, I would recommend using the Turbulent Flow, Spalart-Allmaras (spf) interface. The Creeping Flow Interface that you are using only solves the Stokes equations and therefore neglects the inertial terms necessary to accurately model any flows that aren't vey small Re.

Also, since you are solving in the Time domain, you want to be sure you have allowed the solution to develop in time long enough for you to reach a steady flow state. I see you are solving for a period of 1 second total. I'd bet that this is why your lift coefficient hasn't stabilized yet.

You can also consider using the Stationary form of the equations although these can be more difficult to solve for some flows.

For the future - if you'd like to post *.mph files. Before saving, click Edit>Clear Meshes and Edit>Clear Solutions. This will reduce your file size enough to post it.

Best regards,
Josh Thomas
Ting- For flows with these Reynolds numbers, I would recommend using the Turbulent Flow, Spalart-Allmaras (spf) interface. The Creeping Flow Interface that you are using only solves the Stokes equations and therefore neglects the inertial terms necessary to accurately model any flows that aren't vey small Re. Also, since you are solving in the Time domain, you want to be sure you have allowed the solution to develop in time long enough for you to reach a steady flow state. I see you are solving for a period of 1 second total. I'd bet that this is why your lift coefficient hasn't stabilized yet. You can also consider using the Stationary form of the equations although these can be more difficult to solve for some flows. For the future - if you'd like to post *.mph files. Before saving, click Edit>Clear Meshes and Edit>Clear Solutions. This will reduce your file size enough to post it. Best regards, Josh Thomas
Ting Lau
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago Apr 4, 2013, 12:05 a.m. EDT
Hi Josh,

Thanks for your tip! One of the challenges that I am facing is that my professor do not want to purchase the CFD module where I can use the turbulent flow. He recommend using Laminar solver...I contacted COMSOL if I can purchase a student license but they required the payment coming from the school instead of a student. Therefore, the tool that I have are the outdated Multiphysics and MEMS module that the school owns...

For my model, I increased the times study but it causes the Lift coefficient to keep decreases as time increases?

Would you know what's wrong with my model?

I am able to attached the file now.
Thanks for your tip!

Ting
Hi Josh, Thanks for your tip! One of the challenges that I am facing is that my professor do not want to purchase the CFD module where I can use the turbulent flow. He recommend using Laminar solver...I contacted COMSOL if I can purchase a student license but they required the payment coming from the school instead of a student. Therefore, the tool that I have are the outdated Multiphysics and MEMS module that the school owns... For my model, I increased the times study but it causes the Lift coefficient to keep decreases as time increases? Would you know what's wrong with my model? I am able to attached the file now. Thanks for your tip! Ting
Ivar KJELBERG COMSOL Multiphysics(r) fan, retired, former "Senior Expert" at CSEM SA (CH)
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago Apr 4, 2013, 3:11 a.m. EDT
Hi

I have taken a look, nice model, at these slow speeds it's indeed interesting to see what is going on, and the turbulent model needs tweaking to solve, more than what I have time to.

I have run it for your case and once for nonstokes flow in-compressible and compressible, it gives some differences during flow build up but then it stabilises, you have the plots here, now I leave it to you to explain the differences between the cases ;)

I'm using a time stepping of: 2^{range(1,0.5,6)} range(80,20,500)

I'm using 4.3a

--
Good luck
Ivar
Hi I have taken a look, nice model, at these slow speeds it's indeed interesting to see what is going on, and the turbulent model needs tweaking to solve, more than what I have time to. I have run it for your case and once for nonstokes flow in-compressible and compressible, it gives some differences during flow build up but then it stabilises, you have the plots here, now I leave it to you to explain the differences between the cases ;) I'm using a time stepping of: 2^{range(1,0.5,6)} range(80,20,500) I'm using 4.3a -- Good luck Ivar
Josh Thomas Certified Consultant
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago Apr 4, 2013, 10:11 a.m. EDT
Ting -

How do you know something is "wrong" with your model?

If the flow characteristics are still changing in time, then this indicates that you have not acheived steady-state. If a problem indeed has a steady-state solution (some natural convection flows for example do not) then you need to extend the time period and allow the flow to stabilize in time. This flow should have a steady-state solution which implies that as you continue in time you will approach this state.

The meaning and physical relevance of the transient portion of your simulation is harder to interpret. Why is it surprising that during this transient "start-up" portion of your simulation that the lift coefficient drops?

You should ask yourself: What am I trying to model? What are you looking to characterize about this airfoill?

I have attached the results from running your simulation with the range settings suggested by Ivar and after 500 seconds there is still some change in the lift and drag but not much. I'd bet that if solved for a longer period things would stabilize even more.

Have you tried solving using the stationary solver?

Regards,
Josh Thomas
AltaSim Technologies
Ting - How do you know something is "wrong" with your model? If the flow characteristics are still changing in time, then this indicates that you have not acheived steady-state. If a problem indeed has a steady-state solution (some natural convection flows for example do not) then you need to extend the time period and allow the flow to stabilize in time. This flow should have a steady-state solution which implies that as you continue in time you will approach this state. The meaning and physical relevance of the transient portion of your simulation is harder to interpret. Why is it surprising that during this transient "start-up" portion of your simulation that the lift coefficient drops? You should ask yourself: What am I trying to model? What are you looking to characterize about this airfoill? I have attached the results from running your simulation with the range settings suggested by Ivar and after 500 seconds there is still some change in the lift and drag but not much. I'd bet that if solved for a longer period things would stabilize even more. Have you tried solving using the stationary solver? Regards, Josh Thomas AltaSim Technologies

Attachments:

Bill
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 8 years ago Nov 22, 2016, 3:38 a.m. EST

Ting -

How do you know something is "wrong" with your model?

If the flow characteristics are still changing in time, then this indicates that you have not acheived steady-state. If a problem indeed has a steady-state solution (some natural convection flows for example do not) then you need to extend the time period and allow the flow to stabilize in time. This flow should have a steady-state solution which implies that as you continue in time you will approach this state.

The meaning and physical relevance of the transient portion of your simulation is harder to interpret. Why is it surprising that during this transient "start-up" portion of your simulation that the lift coefficient drops?

You should ask yourself: What am I trying to model? What are you looking to characterize about this airfoill?

I have attached the results from running your simulation with the range settings suggested by Ivar and after 500 seconds there is still some change in the lift and drag but not much. I'd bet that if solved for a longer period things would stabilize even more.

Have you tried solving using the stationary solver?

Regards,
Josh Thomas
AltaSim Technologies

Dear Josh,
I am facing a problem about the evolution of lift coefficient with time in my model, as in the attachment.
I want to find out when it will come to a steady solution using the lift coefficient as a characterization parameter. I wonder if that is a suitable choice.
I have got a stongly oscillating values for the range(0,0.1,100), which is quite different from what you have posted.
In addition, I used periodic boundary conditions, and I set the initila value of u as 1, which I used to calculate the lift coefficient. I wonder if that is a correct value or should I make evaluation of the flow field to get a mean velocity?
Thank you in advance.
Best Regards,
Bill
[QUOTE] Ting - How do you know something is "wrong" with your model? If the flow characteristics are still changing in time, then this indicates that you have not acheived steady-state. If a problem indeed has a steady-state solution (some natural convection flows for example do not) then you need to extend the time period and allow the flow to stabilize in time. This flow should have a steady-state solution which implies that as you continue in time you will approach this state. The meaning and physical relevance of the transient portion of your simulation is harder to interpret. Why is it surprising that during this transient "start-up" portion of your simulation that the lift coefficient drops? You should ask yourself: What am I trying to model? What are you looking to characterize about this airfoill? I have attached the results from running your simulation with the range settings suggested by Ivar and after 500 seconds there is still some change in the lift and drag but not much. I'd bet that if solved for a longer period things would stabilize even more. Have you tried solving using the stationary solver? Regards, Josh Thomas AltaSim Technologies [/QUOTE] Dear Josh, I am facing a problem about the evolution of lift coefficient with time in my model, as in the attachment. I want to find out when it will come to a steady solution using the lift coefficient as a characterization parameter. I wonder if that is a suitable choice. I have got a stongly oscillating values for the range(0,0.1,100), which is quite different from what you have posted. In addition, I used periodic boundary conditions, and I set the initila value of u as 1, which I used to calculate the lift coefficient. I wonder if that is a correct value or should I make evaluation of the flow field to get a mean velocity? Thank you in advance. Best Regards, Bill

Attachments:

Note that while COMSOL employees may participate in the discussion forum, COMSOL® software users who are on-subscription should submit their questions via the Support Center for a more comprehensive response from the Technical Support team.

Suggested Content