4 Replies Latest reply on Jul 11, 2014 4:04 AM by Carel Kriek

    Simulation jumps to Large Problem Direct Sparse solver

    Carel Kriek

      Hi, I am trying to solve a linear static study with a couple of defined contact sets defined as no-penetration with node-to-surface option checked. About twenty bolt connectors are used in the study. Degrees of freedom are 109,698, Number of nodes are 35,327 and number of elements are 21,933. My problem is twofold:

       

      a. the solver remains on Current task: Solving for contact restraints (it just remains there for hours with no warnings)

      b. The solution jumps automatically to solver type: Large Problem Direct Sparse. (I cannot get it to run the FFEPlus solver or the Direct Sparse solver even if I try to set it that way)

       

      It is very simple geometry basically consisting out of two bolted together flanges. One flange with fixed restraints and the other flange with applied force and torque.

       

      When I terminate the solver and save, I do get some results that look like what I expect but obviously cannot be trusted since the solver was interrupted.

       

      My machine is a Dell T7500 with 64 GB RAM.

       

      Can anybody help with a) and b) above? I will really appreciate!

       

      Regards.

       

      Carel.

       

      PS. See attached file for the message I get after running the study.

       

      Message was edited by: Carel Kriek

        • Re: Simulation jumps to Large Problem Direct Sparse solver
          Jared Conway

          time to go into troubleshooting mode

          go to bonded contact, does it run? if not, you're going to have a tough time getting no pen to run

          if that works, go to a lower preload or make sure that you have enough restraints in your model

            • Re: Simulation jumps to Large Problem Direct Sparse solver
              Carel Kriek

              Thanks Jared,

               

              Yes, I removed all no-penetration contacts and make all bonded. Removed all bolted connectors. Refined mesh very much (15 fold). The simulation ran literally in an eyewink! It would still not allow me to run the FFEPlus solver even if I select it. During solving it shows the Large Problem Direct Sparse Solver. What could be the reason for this?

               

              I understand the lower preload refers to the torque I specified for the bolted connectors?

               

              Will try the lower preload and see what happens. (I had 25 Nm - will lower to 5 Nm to start off with?)

               

              -------------------------------------------------------------------------------------------------------------------------------

              Update: I removed all no-penetration contacts and all bolt connectors and ran the simulation with a coarser mesh. (Simulation ran quick). Then I compiled a Submodel study and add the relevant no-penetration contacts and bolted connectors to the much reduced model (only two parts left). Remeshed to a fine mesh and ran the submodel study ( 1,5 M DoF). The solution came within 30 minutes and I am very happy with the results!

               

              However I still seem not to be able to control/choose which solver to use.

               

              Attached is a Picture of stress results.

               

              Regards.

               

              Carel.

               

              Message was edited by: Carel Kriek