Hi everyone,
I’m trying to learn how to use the colloidal isotherm, and my test simulation is failing to converge. The .h5 file for my test is attached, along with an associated notebook. I’m using isotherm parameters comparable to those in the BindingModels.cpp test case. Could you please direct me on the colloidal isotherm’s appropriate usage/let me know if I’m missing anything? (I don’t believe this is included in the user documentation yet.)
Thank you!
Chase
colloidal_test.h5 (40.3 KB)
test_colloidal_isotherm_no_imports.ipynb (16.0 KB)
Hey Chase,
you are right, the model is still not supported officially. But actually I’ve been working on the documentation in the last couple of days. If you want to have a look, checkout the PR and build the docs locally (see here for instructions).
I’m actually awaiting @vijesh’s review. Maybe he can also help you how to set up the model.
I haven’t looked at your files, but the analytic Jacobian of the model is still incorrect.
If you haven’t done so already, please try automatic differentiation by setting USE_ANALYTIC_JACOBIAN = 0.
Thank you, Johannes and Sam. I had set USE_ANALYTIC_JACOBIAN = 0, and my usage appears consistent with the documentation, but clearly I could be missing something. I will let you know if I find the problem.
I suggest contacting @vijesh (PostDoc in Delaware in Lenhoff’s lab). He is successfully using this model and may be able to provide some support.
Apparently the magnitude of some parameters precluded solver convergence. Attached is a simulation with example parameter values that work (based on my discussion with Vijesh).
test_colloidal_isotherm_no_imports.ipynb (27.0 KB)