Newbie needs advice for a school project!
- Mark
- Topic Author
- Offline
- New Member
-
Less
More
- Posts: 2
- Thank you received: 0
11 years 10 months ago #7099
by Mark
Newbie needs advice for a school project! was created by Mark
Hi everyone.
I am a college student working on an honors project but my school doesn't have anyone or any resources to help me so I was hoping someone here can point me in the right direction.
I'm new to CFD AND FEA, trying to self-teach myself with tutorials and youtube videos, and was hoping someone with experience could help me understand a couple of conceptual questions about CAELINUX, Salome, and OpenFoam.
My project is rather simple.
I drew a model in Salome of a small object that is supposed to represent a small rock inside a large rectangular box with open ends, where one end represents the inlet, and the other the outlet. The large box is supposed to represent the outside environment.
I simply want to run several CFD simulations of wind rushing in one side of the box, interacting with the "rock", and then exiting the other side.
This whole setup is simply to find out how much wind it would take to lift-up or make the "rock" move or tumble. The geologic term would be called saltation for those interested.
What I DO understand is that it isn't very hard to run this simulation and see the fluid dynamics of the wind interacting with the rock and then exiting the system. I mean the simulation will show pressure gradients and what-not, however that isn't the problem.
What I DON'T understand is how will I know if the rock will move or not?
For example, if someone drew a simple airfoil and placed it in a space with boundary conditions, how do they know that their airfoil will lift up and fly?
I mean, the software can't physically show an animation of the airfoil being lifted up and taking flight can it?
And if that assumption is correct, then it seems to me that someone testing an airfoil by drawing it in Salome, then meshing, and running a simulation in SimpleFoam would only have the pressure gradients on which to base their findings right?
And if that assumption is correct, how do they take those numbers and declare then that their airfoil would actually fly?
I guess another way to look at what I'm trying to understand would be like doing a stress simulation on a pipe or something to see what its breaking point would be. Does the software actually show the pipe breaking in an animation or does the person running the simulation somehow extrapolate the data and then somehow declares that the pipe would break at certain points under certain conditions?
Sorry this is so long, but as you can see, I do not have any mentors to help or guide me.
Perhaps someone with the experience could give me a brief synopsis on how the pros go about doing simulations like this so I can understand the limits of the software and/or the requirements needed of someone running a simulation and their ability to be able to pluck the data they need to answer their questions from the finished simulation?
If anyone can help me learn about these questions or suggest a better way for me to complete my project, I would be very grateful!
Thanks in advance for all your help CAELINUX Forum users!
-MJ
I am a college student working on an honors project but my school doesn't have anyone or any resources to help me so I was hoping someone here can point me in the right direction.
I'm new to CFD AND FEA, trying to self-teach myself with tutorials and youtube videos, and was hoping someone with experience could help me understand a couple of conceptual questions about CAELINUX, Salome, and OpenFoam.
My project is rather simple.
I drew a model in Salome of a small object that is supposed to represent a small rock inside a large rectangular box with open ends, where one end represents the inlet, and the other the outlet. The large box is supposed to represent the outside environment.
I simply want to run several CFD simulations of wind rushing in one side of the box, interacting with the "rock", and then exiting the other side.
This whole setup is simply to find out how much wind it would take to lift-up or make the "rock" move or tumble. The geologic term would be called saltation for those interested.
What I DO understand is that it isn't very hard to run this simulation and see the fluid dynamics of the wind interacting with the rock and then exiting the system. I mean the simulation will show pressure gradients and what-not, however that isn't the problem.
What I DON'T understand is how will I know if the rock will move or not?
For example, if someone drew a simple airfoil and placed it in a space with boundary conditions, how do they know that their airfoil will lift up and fly?
I mean, the software can't physically show an animation of the airfoil being lifted up and taking flight can it?
And if that assumption is correct, then it seems to me that someone testing an airfoil by drawing it in Salome, then meshing, and running a simulation in SimpleFoam would only have the pressure gradients on which to base their findings right?
And if that assumption is correct, how do they take those numbers and declare then that their airfoil would actually fly?
I guess another way to look at what I'm trying to understand would be like doing a stress simulation on a pipe or something to see what its breaking point would be. Does the software actually show the pipe breaking in an animation or does the person running the simulation somehow extrapolate the data and then somehow declares that the pipe would break at certain points under certain conditions?
Sorry this is so long, but as you can see, I do not have any mentors to help or guide me.
Perhaps someone with the experience could give me a brief synopsis on how the pros go about doing simulations like this so I can understand the limits of the software and/or the requirements needed of someone running a simulation and their ability to be able to pluck the data they need to answer their questions from the finished simulation?
If anyone can help me learn about these questions or suggest a better way for me to complete my project, I would be very grateful!
Thanks in advance for all your help CAELINUX Forum users!
-MJ
- Chris M
- Offline
- New Member
-
Less
More
- Posts: 17
- Thank you received: 3
11 years 10 months ago #7100
by Chris M
Replied by Chris M on topic Re: Newbie needs advice for a school project!
Hey there,
So from what I understand you are running a simulation of airflow over a rock and would like to know when the rock will lift up. You plan on running several simulations of the wind going at different speeds to see when it will tumble. But you don't know what to look for in these simulations.
So to try and help lets go over what you may be doing. I think you will be doing a "steady-state" run. So it doesn't change with time, its just a snapshot of the physics at one point in time assuming that the condition are constant at all points in time. The opposite of this being time-dependent where the conditions change with respect to time. Doing steady state will give you a snapshot and you can analyse the physics after the simulation has converged to a solution. That means that after the simulation has iterated over the Navier-Stokes numerical equations enough times that the error between each simulation has lowered enough to an acceptable threshold.
The analysis comes in now. You want to look at the pressure difference between the upper and lower surfaces of your rock. The difference in pressure will determine if the rock will tumble or not. After you find the pressure difference you use that to find if the rock is producing lift. This is from NASA's website. Basically you are looking for the force. Force is pressure times area. You should know the area of the rock, or the half that you consider the "top" of the rock. And you should have the pressure for that area. If you did the simulation correctly then you should have meshed the surface of the rock to do analysis on its surface. There should be cells, triangles or squares, on the surface and each cell will have a value for the pressure. Each cell will also have a value for its own area so just multiply the two together. Then perform an integration over the surface and get the total force.
Now to determine if the rock is tumbling or not. First you need to know all the forces acting on the rock. This includes the weight acting downward and the drag force or friction force. You need the coefficient of friction static because the rock is initially not moving. Use this site for reference. Figure out the forces acting on the rock before the wind hits, figure out the forces acting on it after the wind hits by using the pressure and areas on the rock, sum all the forces in like directions together, then you'll know if the rock moved if the force in same direction of the wind is positive.
I hope this helps.
So from what I understand you are running a simulation of airflow over a rock and would like to know when the rock will lift up. You plan on running several simulations of the wind going at different speeds to see when it will tumble. But you don't know what to look for in these simulations.
So to try and help lets go over what you may be doing. I think you will be doing a "steady-state" run. So it doesn't change with time, its just a snapshot of the physics at one point in time assuming that the condition are constant at all points in time. The opposite of this being time-dependent where the conditions change with respect to time. Doing steady state will give you a snapshot and you can analyse the physics after the simulation has converged to a solution. That means that after the simulation has iterated over the Navier-Stokes numerical equations enough times that the error between each simulation has lowered enough to an acceptable threshold.
The analysis comes in now. You want to look at the pressure difference between the upper and lower surfaces of your rock. The difference in pressure will determine if the rock will tumble or not. After you find the pressure difference you use that to find if the rock is producing lift. This is from NASA's website. Basically you are looking for the force. Force is pressure times area. You should know the area of the rock, or the half that you consider the "top" of the rock. And you should have the pressure for that area. If you did the simulation correctly then you should have meshed the surface of the rock to do analysis on its surface. There should be cells, triangles or squares, on the surface and each cell will have a value for the pressure. Each cell will also have a value for its own area so just multiply the two together. Then perform an integration over the surface and get the total force.
Now to determine if the rock is tumbling or not. First you need to know all the forces acting on the rock. This includes the weight acting downward and the drag force or friction force. You need the coefficient of friction static because the rock is initially not moving. Use this site for reference. Figure out the forces acting on the rock before the wind hits, figure out the forces acting on it after the wind hits by using the pressure and areas on the rock, sum all the forces in like directions together, then you'll know if the rock moved if the force in same direction of the wind is positive.
I hope this helps.
The following user(s) said Thank You: Mark
- Mark
- Topic Author
- Offline
- New Member
-
Less
More
- Posts: 2
- Thank you received: 0
11 years 10 months ago #7101
by Mark
Replied by Mark on topic Re: Newbie needs advice for a school project!
Thanks Eezyville!
This is awesome information and definitely enough to chew on for a week or so to get me going in the right direction.
With respect to this excerpt from your post:
"There should be cells, triangles or squares, on the surface and each cell will have a value for the pressure. Each cell will also have a value for its own area so just multiply the two together. Then perform an integration over the surface and get the total force."
While I still need to run a simulation, however when that simulation is done and I pick a "static" moment in time in which the wind is acting on and around the rock, how exactly do I find the areas of each triangular cell? Do I zoom in on them or is this data listed as a time-stamped text file or something?
I realize that question could be a sort of Pandora's Box so to speak and there might be multiple ways to answer based on the specific set-ups of the simulation run, but do you know of any tutorials or videos that I might be able to study and then directly apply those methods to my situation so that I may help reduce this crazy learning curve I am faced with?
Also, another mention, I'm a little confused about how you say to "integrate" over the surface. The only integration I have done thus far in my academic career is to take a various function and integrate that in order to find the area under that function's specific curve. I guess what I'm confused about is that your post made it sound like the software will have already calculated not only the area within each triangular cell for a specific moment in time, but also the pressure at that moment. How does one go about integrating a scenario like this without a "typical" function like I mentioned?
At the moment, the only solution I can think of is some sort of brute-force arithmetic where I would physically gather the area and pressure data from each cell "in the red so to speak" and then multiply and add them all together. Now of course I realize that this methodology is coming from someone, "me", with a rudimentary understanding of the process, but I thought I would just mention how I am thinking of this problem so that you may offer a better way of addressing it.
Guess this all might be answered if perhaps you have any leads on an airfoil tutorial or videos.
Anyway, thanks again for your effort and time in helping me! Much appreciated!
-MJ
This is awesome information and definitely enough to chew on for a week or so to get me going in the right direction.
With respect to this excerpt from your post:
"There should be cells, triangles or squares, on the surface and each cell will have a value for the pressure. Each cell will also have a value for its own area so just multiply the two together. Then perform an integration over the surface and get the total force."
While I still need to run a simulation, however when that simulation is done and I pick a "static" moment in time in which the wind is acting on and around the rock, how exactly do I find the areas of each triangular cell? Do I zoom in on them or is this data listed as a time-stamped text file or something?
I realize that question could be a sort of Pandora's Box so to speak and there might be multiple ways to answer based on the specific set-ups of the simulation run, but do you know of any tutorials or videos that I might be able to study and then directly apply those methods to my situation so that I may help reduce this crazy learning curve I am faced with?
Also, another mention, I'm a little confused about how you say to "integrate" over the surface. The only integration I have done thus far in my academic career is to take a various function and integrate that in order to find the area under that function's specific curve. I guess what I'm confused about is that your post made it sound like the software will have already calculated not only the area within each triangular cell for a specific moment in time, but also the pressure at that moment. How does one go about integrating a scenario like this without a "typical" function like I mentioned?
At the moment, the only solution I can think of is some sort of brute-force arithmetic where I would physically gather the area and pressure data from each cell "in the red so to speak" and then multiply and add them all together. Now of course I realize that this methodology is coming from someone, "me", with a rudimentary understanding of the process, but I thought I would just mention how I am thinking of this problem so that you may offer a better way of addressing it.
Guess this all might be answered if perhaps you have any leads on an airfoil tutorial or videos.
Anyway, thanks again for your effort and time in helping me! Much appreciated!
-MJ
- Chris M
- Offline
- New Member
-
Less
More
- Posts: 17
- Thank you received: 3
11 years 10 months ago - 11 years 10 months ago #7102
by Chris M
Replied by Chris M on topic Re: Newbie needs advice for a school project!
The cell area should be calculated by the software that you used to mesh the rock. It has to do that because it needs to know if the mesh was done right. You don't want to have cells with negative areas now do you? If you have a negative area then it may mean that the cell normal vector is pointing into the wrong direction and the calculations will be messed up on that cell. Whatever program you used to mesh the cell should know its area and the CFD program that you use will also know the area. You have to look through the documentation to find it. But I may have a simpler solution which I will explain later.
When I was talking about integration I was speaking on terms of numerical integration. You may want to do some research on how to do it. I admit that I haven't been able to work on a problem like this often even though it sounds relatively simple. I'm also not sure what resources you have available from your school. You could have commercial CFD software that will do the work for you.
Now the simple solution would be to learn how to use a post-processing program like Tecplot or Paraview to analyze your solution. I have experience with Tecplot and I know that if you were to run a solution using something like Fluent and import the data into Tecplot you will have what you need for this problem. You will have the pressure and area values for each cell. Then write a custom function to calculate the force, which is easy to do. You can also perform numerical integration using Tecplot or have Tecplot calculate the force for you. Get a trial version of it from their website or you can learn how to use Paraview, which is free. But I don't know how to use Paraview yet.
For more help please visit CFD-Online. They are a great resource and they have something for education for students like you.
When I was talking about integration I was speaking on terms of numerical integration. You may want to do some research on how to do it. I admit that I haven't been able to work on a problem like this often even though it sounds relatively simple. I'm also not sure what resources you have available from your school. You could have commercial CFD software that will do the work for you.
Now the simple solution would be to learn how to use a post-processing program like Tecplot or Paraview to analyze your solution. I have experience with Tecplot and I know that if you were to run a solution using something like Fluent and import the data into Tecplot you will have what you need for this problem. You will have the pressure and area values for each cell. Then write a custom function to calculate the force, which is easy to do. You can also perform numerical integration using Tecplot or have Tecplot calculate the force for you. Get a trial version of it from their website or you can learn how to use Paraview, which is free. But I don't know how to use Paraview yet.
For more help please visit CFD-Online. They are a great resource and they have something for education for students like you.
Last edit: 11 years 10 months ago by Chris M. Reason: typos
The following user(s) said Thank You: Mark
- Volker Claus Falk
- Offline
- New Member
-
Less
More
- Posts: 16
- Thank you received: 2
11 years 10 months ago - 11 years 10 months ago #7108
by Volker Claus Falk
Replied by Volker Claus Falk on topic Re: Newbie needs advice for a school project!
Hi MJ, Hi Eezyville, Hi all,
coming from civil with also only a little experience in fluid dynamics
I just see your prob finally as the question of a free motion of a more-or-less
rigid body (the rock) in a time and place -dependend more-less compressible
fluid field (the air acting as wind).
Fluid computation in FEA is big stuff and could have been done with PCs
not too long time ago ... before the researchers used big CRAY computers or
in older times SIEMENS or UNIVAC or alike.
This so called CFD needs - also for a small sample - many small time steps to calculate
the changing of the fields and the moverment of the body within this small time steps.
As a result you get a huge amount of data out of which an appropriate post-processor
program could calculate and show the information you need (e.g. the deflection way
of the center of your body or maybe also a animation of the bodies movement in the fluids
stream or alike).
I do not know anything about OpenFoam but I found on CA website some link that maybe
could be interesting for you:
and a there a special link for combining SALOME with CA
My remarks and the links cannot be the cookbook you maybe are looking for,
but IMHO the discussion on the CA forum shows a little about the basical
problems in the modelling of nonlinear non-steady state fluid dynamics in FEA.
Wish you that you will find anywhere a small simulation example (balls in water flux or alike)
where you could study the stuff - but thats a time consumimg field that you cannot cover within some days or even weeks.
All the Best - Volker
coming from civil with also only a little experience in fluid dynamics
I just see your prob finally as the question of a free motion of a more-or-less
rigid body (the rock) in a time and place -dependend more-less compressible
fluid field (the air acting as wind).
Fluid computation in FEA is big stuff and could have been done with PCs
not too long time ago ... before the researchers used big CRAY computers or
in older times SIEMENS or UNIVAC or alike.
This so called CFD needs - also for a small sample - many small time steps to calculate
the changing of the fields and the moverment of the body within this small time steps.
As a result you get a huge amount of data out of which an appropriate post-processor
program could calculate and show the information you need (e.g. the deflection way
of the center of your body or maybe also a animation of the bodies movement in the fluids
stream or alike).
I do not know anything about OpenFoam but I found on CA website some link that maybe
could be interesting for you:
http://www.code-aster.org/forum2/viewtopic.php?id=14400
and a there a special link for combining SALOME with CA
http://www.aero.polimi.it/freecase/?download=RomanelliSerioli_MSCThesis_Presentation.pdf
My remarks and the links cannot be the cookbook you maybe are looking for,
but IMHO the discussion on the CA forum shows a little about the basical
problems in the modelling of nonlinear non-steady state fluid dynamics in FEA.
Wish you that you will find anywhere a small simulation example (balls in water flux or alike)
where you could study the stuff - but thats a time consumimg field that you cannot cover within some days or even weeks.
All the Best - Volker
Last edit: 11 years 10 months ago by Volker Claus Falk. Reason: typo
The following user(s) said Thank You: Mark
Moderators: catux
Time to create page: 0.195 seconds