Posted Tue, 08 Mar 2016 20:59:15 GMT by zhengq
the HELP with FEFLOW says:

[i]The Direct solution strategy provides the advantage of an exact solution of the equation system without any residual error due to a solver iteration. This might come at the expense of longer runtimes and higher memory demand, depending on problem setup and available hardware.[/i]

But when I choose direct solvers, I got the following info:

                  No convergence in iteration loop at time: 1.000000e-03 with remaining error: 2.990310e+06 greater than  1.000000e-03

it is contradictory. Any comments and help are appreciated.

Posted Tue, 08 Mar 2016 21:44:45 GMT by Denim Umeshkumar Anajwala
There are both an outer and an inner iteration. The outer one is executed in all cases where you have a non-linear equation system to solve (unconfined aquifers, unsaturated simulation, density-dependent simulation, ...). The inner one is the solver iteration, and this one you do not have in case of a direct equation-system solver. The non-convergence message refers to the outer iteration that is controlled by the error criterion on the Numerical Parameters page of the Problem Settings dialog.

You must be signed in to post in this forum.