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.
Flow Boiling Simulation Convergence Problems
Posted 2026/04/17 9:36 GMT-4 Fluid & Heat, Computational Fluid Dynamics (CFD), Studies & Solvers Version 6.4 1 Reply
Please login with a confirmed email address before reporting spam
I am trying to adapt the extended RPI flow boiling model to my own expiramental setup where i have liquid nitrogen flowing through two channels in a metal test piece, with known heat flux applied to the bottom side. I have replicated the exact setup as the published application. The convergence has failed consistently. I have tried auxillary sweep on the boiling and heat flux, running fully coupled and segregated studies, putting a solver limit on the volume fraction, to no result.
This is how it has been failing:
Undefined value found. - Detail: Undefined value found in the stiffness matrix. There are 111 equations giving NaN/Inf in the matrix rows for the variable comp1.T. at coordinates: (0.0369032,0.029986,0.0268232), (0.0377371,0.0311923,0.026833), (0.0370727,0.0323953,0.0268382), (0.0348952,0.0336535,0.0268406), (0.0341695,0.0347622,0.0268414), ...
and similarly for the degrees of freedom, NaN/Inf in the matrix columns.
here is the application, and screenshots of my own model in an experiment that converged using water, as a reference: https://www.comsol.com/model/subcooled-nucleate-boiling-with-extended-rpi-model-113081
Any advice for convergence with phase change?
The file is too large to send here, but happy to provide more detail to anyone who is willing to help.
thanks!
Attachments:
Please login with a confirmed email address before reporting spam
Consider cleaning your mesh and studies results, then compacting history to significantly reduce the file size and then uploading here. Usually NaN/Inf results from a extremely low mesh quality or not properly setting boundary conditions, but also can be from something else.
