Quantcast
Channel: Latest Discussions - COMSOL Forums
Viewing all articles
Browse latest Browse all 5773

Convergence issue when using a small constant (May have reached a singularity)

$
0
0

Hello I am solving a 2D transient Poisson and Nernst-Planck equations on two domains. The initial conditions are different in two domains. In the Poisson equations (There are 2 Poisson equations), on the right hand side I have a constant (epsilon) which is defined in Parameters. This is a small constant (Please see the attached photos) I can solve the problem when it is equal 8.85 * (10^(-8)) but the correct value of this constant is : 8.85 * (10^(-10)) . When I use this value the problem does not converge and I get this error: comsol Repeated error test failures. May have reached a singularity I am using PARDISO solver and I also changed the solver setting as recommended in this link : www.comsol.com/support/knowledgebase/1172/ I also decreased the relative tolerance in different steps but none of these tricks was helpful. I am pretty sure there should be a way to solve this problem for epsilon = 8.85 * (10^(-10)) The thing is I can not change the value of initial or boundary conditions or any constants in my model. For example if I multiply the initial conditions by 0.001 this problem converges perfectly but I do not want to change my initial conditions.** I was hoping if using a preconditioner or changing the solver or mesh or WHATEVER could resolve my issue. ** I would really appreciate your time taking a look at my model and help me get around this problem. The file of my model and two screenshots are attached. Thanks in advance for your support.


Viewing all articles
Browse latest Browse all 5773

Trending Articles



<script src="https://jsc.adskeeper.com/r/s/rssing.com.1596347.js" async> </script>