Topology Optimisation of Trusses problem
Altair EmployeeHii,
I am doing Topology Optimisation of Truss (Benchmark Problem).I am not getting proper results of topology as expected.Please do suggest me what to do for proper analysis & optimisation.I am unable to find out my mistake.I am attaching my files
Answers
-
HI,
Can you give us any reference of what you are expecting (benchmark problem)?
0 -
0
-
HI,
Can you use displacement and mass/volume as responses instead of compliance and VF?
<?xml version="1.0" encoding="UTF-8"?>
0 -
Sir, I am learner, I will do whatever you suggest me. can you give me the file in which you have done.. I will understand from that, what procedure to adopt & how to apply.
0 -
0
-
Hii Sir,
Thanks for your valuable advice for Topology Optimization of Truss.But If I do the same problem for size Optimization the error message appears
'ERROR # 568 ***
There are no topology nor referenced design variables.Use the ANALYSIS keyword to perform analysis only.'Please do the needful I am attaching my files.
0 -
Hi @NINAD
What type of optimization do you want to carry?
In your last post, you have DSIZE variable with no design parameters. Please refer to size optimization tutorials if you are interested in sizing
0 -
Sir, I have followed the tutorial & do as it taught me. But what to do in generic Relationship in my case I am not clarify in it.what is design parameter. I want to do size optimization in my problem, reducing the size of element is my aim.
0 -
HI,
Here is a sample file of your model. Please change the parameters of your choice and run the analysis.
0 -
Sir,
After Size optimization ,the resulting window only shows Displacement & stress in Result type tool.It shows Analysis type of window as such we do in analysis having iteration only 2 in no.
<?xml version="1.0" encoding="UTF-8"?>
0 -
Sir,
In my software the toggle option for thickness does not appears in generic relationship of design variable of size optimization.I don't know why ?what to do then.?
<?xml version="1.0" encoding="UTF-8"?>
0 -
HI Ninad,
I see that you are using ROD element and you can use the area as a design variable.
To use thickness use a different element type.
Also, at the end of optimization, a .prop file will be written with new area/thickness/etc... information
0 -
Sir,
I am trying my best to get results but everytime a error or something new appears as described above.Please solve the problem in size optimization.I have declared everything according to the tutorials.I want to reduce the area of element i.e. rod.I am attaching my file for you
0 -
Sir,
Please do the favor. please do resolve the problem asap.
0 -
Try this:
You are trying to minimize the volume and cut the displacement by almost 3.5 times.
May not be possible. Try a different approach.
<?xml version="1.0" encoding="UTF-8"?>
0 -
Sir,
I have done what i can do.But the same error exists or the result window shows the results as shown above,only displacement & stresses.Not doing size optimization in the truss file.Please suggest some remedy or solved the model by yourself.It will be your very kind.
0 -
Hi Ninad,
Please share what the end objective.
0 -
Sir,
The end objective is to minimise the size of rod for the apply loads.Hence the truss has reduced size of members i. e. rod element.
0 -
Sir,
Please do the favour. solve the problem
0 -
Hi Ninad,
as said you already have optimal design and beyond reducing the mass will lead to failure.
You can see through the error between the iterations that the load case is infinite compliance.
0 -
Hi everyone,
I have tried to solve this benchmark problem. But i am not able to obtain the contour plot of the element densities. So how can I obtain these values in hyperview?
If I understand correctly, the minimum member area is already specified in this problem. Therefore the solver should be able to reduce the member area to zero for all the members that are not structural 'optimal' and this should lead to the benchmark solution of the 10 bar truss problem. How can I properly set-up this 10 bar truss problem in order to check my optimization set-up by comparing it to this benchmark solution. Your feedback is welcome.Regards,
Michael0 -
Dear optistruct users,
In the meantime, I have tried to set-up the 10 bar truss problem by myself. Unfortunately I am unable to obtain the same results as the benchmark solution.
I used the following benchmark solution in order to verify my method. My problem definition can be found in the following file:
<?xml version="1.0" encoding="UTF-8"?>
The values that I obtain are in the range but they are not the same. Can someone check my model or set-up?
Any feedback would be appreciated.
In the following link there is a describtion for a 10 bar truss optimization in optistruct, but I couldn't locate the file which belongs to it.
Regards,
Michael
0 -
0
-
Dear Prakash,
Do you have any updates about my model? Or can you provide me the example file which is mentioned in the link above?
Regards,Michael
0 -
Hi @michaelv
I tried to download the file but looks like the file is corrupted.
I didn't get a chance to model the same from scratch. I will try to do it ASAP!!
0 -
Dear Prakash,
Thanks for your reply, I think something went wrong while uploading the file.
Because the file had no size at all. Here you can find the model. Thanks for the help.
Regards Michael
0 -
0
-
Dear Prakash,
Thanks for sharing the file. I have changed the support conditions and lower bounds for the cross sectional area.
Now the solution is close to the solution of the benchmark problem. The difference is probably due to the fact the these results were obtained by another method.
The deviation between the two objective values is 0.4%. I will further investigate your model to check my mistakes.
Now I only have one question: Why do you have set the boundary condition only at the lower node (Node 6 in the picture), because this cannot ensure stability.
This lower node is only constrained in the X and Y direction (DOF12) and therefore the structure will rotate around the Z-axis. When I run the analysis I do not see anything strange.
Is this true or do I miss something?Kind regards,
Michael0 -
Hi @michaelv
This how the benchmark problem is formulated.
You can also constrain the other node and this gives a closer result as well.
0 -
Allright,Now I have defined the problem by myself in exactly the same manner as in your .fem file.
The strange thing is that the member size are optimized up to the specified lower limit and then the optimization is ended.
This should mean that there is something wrong with my constraints, But I cannot see what is wrong. Could you please have a look on my model.
Because I would like to set-up these kind op problems by myself. Thanks in advance! Here is the .fem file.Regards, Michael
0
