I recently discovered this useful tool.
For the preliminary case, I was trying to implement a simple linear elastic uniaxial extension optimization problem using ABAQUS/Explicit. The aim is to estimate the optimum material properties.
I have written the code and I need to compare the displacement field with that of experimental data.
The code, however, is converging after only three iterations and showing the initial estimate of material properties to be the optimum values.
However, based on the experimental data, the optimum values should be E = 300 GPa and Poisson's ratio = 0.3.
I'm attaching the complete code for reference. I would be grateful if someone could look into the matter.
Link for files: https://drive.google.com/drive/folders/1oQJ0-pfUHXnaFDtiQPEJjpBGa2Cdl1GQ?usp=sharing
Regards
Abhilash
Dear Abhilash,
Everything seems ok in the code that you are running. However, from the results of the code it seems that the optimization algorithm cannot converge, probably due to inappropriate values assigned to the various tolerances. For this purpose I have added in the main code some tolerance definitions in which I have asssigned nondefault values as follows:
options.TolX=1e-7; % default value is 1e-10 options.FinDiffRelStep=1e3*sqrt(eps); % default value is sqrt(eps)
Also, since the constraint function is actually not affecting the optimization procedure, I have replaced the following command:
[H,fval,exitflag] = fmincon(@const,H0,[],[],[],[],lb,ub,@confun,options)
with the following:
[H,fval,exitflag] = fmincon(@const,H0,[],[],[],[],lb,ub,[],options)
Please try to run the main code according to the above modifications and let me know if convergence is achieved.
Best regards!