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.
Fitting model to experiment with optimization module when the solution to a parameter should be achiveved with gradually relaxing a nonlinearity suppression
Posted 2017年4月20日 GMT-4 11:36 Optimization Version 5.2a 0 Replies
Please login with a confirmed email address before reporting spam
I have the following problem. I would like to fit a model of an electrolysis cell to experimental current-voltage-curves (CVCs) by finding some effective parameter values. The bias voltage is a global parameter in my model, and in the experimental data the output current is given as a function of bias voltage.
The problem is, that the model contains some very strong nonlinearities, and thus the solution is always achieved by multiplying the nonlinearities with a damping parameter, starting at a value of 10^-10 and gradually relaxing it to 1 in an auxiliary sweep. When I use the optimization module, in the very first step it relaxes the damping, but then, when it would try to calculate the model output for the bias voltage values given in the experimental data, it always wants to go from a damping value of 1, and of course, the model doesn't converge.
Can anybody help me at least in achieving that the optimization solver ramps the damping parameter for all bias voltage values?
In my opinion the cheapest solution would be to relax the damping, ramp the voltage, get the CVC (integral of the current density on the boundary calculated for all bias voltage values) and interpolate between the points of the calculated CVC for the experimental values. Might this be achievable somehow, too?
Hello Lazar Kovacs
Your Discussion has gone 30 days without a reply. If you still need help with COMSOL and have an on-subscription license, please visit our Support Center for help.
If you do not hold an on-subscription license, you may find an answer in another Discussion or in the Knowledge Base.