I am trying to optimize hyperelastic Yeoh C_10, C_20, C_30 parameters and the g_i and characteristic times composing a 5-term Prony series, for a total of 13 involved parameters ( the D_i components are assumed equal to zero).
I have got three different tensile tests performed at 0.1 mm/s, 1 mm/s and 10 mm/s, and I'm using the data matching feature to fit these curves with results from analogous simulation performed in Abaqus. The Isight scheme is shown in the image I've attached.
The optimization technique I've chosen is the Hooke-Jeeves one, and the objective functions to be minimized are the sums of the square differences linked to each set of parameters.
Stopping criteria are so that optimization is stopped when all the objective functions are below 0.1.
The fitting is performed considering displacement and reaction force target data, and a series of boundaries linked to the involved parameters is introduced (so that the g_i terms are between 0 and 1, and the relaxation times are positive).
I've also added a constraint on the sum of the g_i parameters (computed by the calculation tool), which must not be above 1.
The problem is that the optimization cycle stops at a certain time, due either to convergence issues or to errors linked to the sum of viscoelastic deviatoric coefficients (which is said to be above one, despite the constraint I've imposed on it).
I've tried to decrease the minimum time step size and change the parameters boundaries to see if something changed, but the results are always the same: the optimization task doesn't succeed in any case.
Since I'm a newbie and there is not much documentation describing how boundaries, constraints and optimization options should effectively be set and handled, I don't really know what to do at this point. What can I do to solve this problem? Is there any mistake made in the optimization scheme? Are there some constraints linked to stability or other that I'm not considering? Can anybody help me in some way?
