3 Replies Latest reply on May 8, 2012 3:42 AM by Dave Laban

    FEA will not run

    April Dunham

      I am trying to run a pressure test on a part. I use 100 degree celcius and 500 psi. I select the whole part for the temperature and the inside of my tube for the pressure. Solidworks meshes my part then runs the test. At a low percentage, it tells me my displacement is too large and it is too unrealistic for the test to be ran. I do not know what if I am doing something incorrectly, or if there is something wrong with solidworks. If there are any suggestions, please let me know!

        • Re: FEA will not run
          Jerry Steiger

          April,

           

          I'm not familiar with the SolidWorks Simulation user interface, but I would try running a multi-step analysis with a very small initial step size, so that you can see how the model reacts as you begin to apply the pressure. You can also get some insight by just running a single step with a pressure that will allow it to solve.

           

          I'm a bit confused by your 100 degree Celsius. Are changing both the temperature and the pressure, ramping from, say, 20C to 100C as you ramp from 0 to 500 psi? Or are you trying to model what happens at a steady state 100C? I almost never have to worry about temperature, so I am not familiar with how that is handled.

           

          Jerry Steiger

          • Re: FEA will not run
            Anthony Botting

            Hi April: are you running the linear static study type? if so, I have a hunch you're seeing the "large displacement" warning, which I believe turns-on if strain exceeds something like 2%. You could sort-out which test (or both) are causing this. Try the tests separately. Next you could check with hand calculations. I believe for the tube strain due to pressure (assuming circular cross section) you can write: strain = (pressure*radius)/(E*wall thickness), and for thermal strain = (Coefficient of thermal expansion) * (change in temp). To get the zero-strain reference temperature, you can check the study properties and look for an entry by that name ("zero-strain reference temperature").  If either or both of these are over about 2%, you might have to conduct a nonlinear study as suggested by Jerry - you can ramp-up the loading for either test, separately. Hope that helps.

            • Re: FEA will not run
              Dave Laban

              It's probably also worth checking the material properties - I once lost the x10^-5 from a thermal expansion coefficient so needless to say the geometry blew up pretty spectacularly and I had a large displacement error!