3 Replies Latest reply on Apr 30, 2014 3:08 AM by Jared Conway

    How to fix problem with custom coefficient of thermal expansion?

    Carl Eberly

        I am using Simulation to look at thermal stresses and resulting displacements.  The assembly that I am using for this exercise  consists of  solid metal cylinders axially aligned with a glass cylinder sandwiched between the metal cylinders.  I have successfully gotten good data when the middle cylinder is made of SolidWorks "Glass" (SolidWorks Materials / Other Non-metals / Glass).  However, when I tried to see the effect of using Zerodur (a low expansion ceramic) I had to enter the material properties as a custom material.  I was able to get a good thermal analysis, but the static analysis fails with the error message "Thermal coefficient (Alphx) is not defined for element 596".  Element 596 is probably the first element of Zerodur in the mesh.  A previous post concerning "thermal expansion coefficient bug" quoted solution S-059026 in the Knowledge Base.  Unfortunately as a student I don't have acess to the Knowledge Base.  Does anyone know the work arounds for this problem.

       

      I have checked and double checked the custom material data, and the thermal stress analysis always works with "Glass".

       

      I am using SolidWorks Student Edition 2012-2013 which has SolidWorks Premium and SolidWorks Simulation Premium.

        • Re: How to fix problem with custom coefficient of thermal expansion?
          Jared Conway

          i'm confused

          you applied the material in a thermal analysis and the thermal analysis results are "correct"

          but then you run a static analysis and it complains?

          are you using the same materials between the 2 studies and the same mesh between them?

          what happens if you make brand new studies?

          what happens if you make a thermal > thermal stress analysis of a block with that material?

          what happens when you do a static analysis with a temperature applied to the block with your material applied?

            • Re: How to fix problem with custom coefficient of thermal expansion?
              Carl Eberly

              Hi Jared,

               

              Thankyou for your reply.  The short answer is that I tried again to copy a SolidWorks material and paste into my Custom materials library and this time it worked.  I was able to change the parameters to Zerodur's and it continued to work.  The material I set up yesterday looks to have identical parameters but it continues to not work.  So maybe something got corrupted on the original custom material?

               

              The long answer:

              you applied the material in a thermal analysis and the thermal analysis results are "correct"

              but then you run a static analysis and it complains? YES

              are you using the same materials between the 2 studies and the same mesh between them? YES

              what happens if you make brand new studies? SAME RESULT

              what happens if you make a thermal > thermal stress analysis of a block with that material? SEE BELOW

              what happens when you do a static analysis with a temperature applied to the block with your material applied? SEE BELOW

               

              I followed your advice and simplified the problem to a 2 inch diameter X 1 inch thick cylinder of material with 50 deg C. applied to the central 1" dia on the top and 0 deg C. to the entire bottom surface.  My results with SW's Glass material were:

              Thermal Study

              Glass - Thermal.jpg

              Thermal Stress:

              Glass - Stressl.jpg

               

              For the original Zerodur custom material the results were:

              Thermal Study:

              Zerodur - Thermal.jpg

              but the Zerodur Stress plot gave this:

              Zerodur - Stress.jpg

               

              The two materials are shown below -- the top one is the good one that is working, the bottom one is the bad one.

              Good  Zerodur.jpg

               

              Bad Zerodur.jpg