5 Replies Latest reply on Dec 10, 2014 12:45 AM by Jared Conway

    Scaling material within Simulation?

    Mark Kamps

      Alright, let me explain the situation:


      For a school project I am working with Simulation. I am trying to create a surface-like part, and I want to 'press' a solid plate on top of this 'surface'. I want the assembly to be analysed using FEA. I will add the files as an attachment.

      With this FEA I want to extract the contact and normal forces applying on the surface. I want to extract the deformed surface to calculate the 'mean' of the surface, to see how much it has been deformed. With these forces and deformations I want to do calculations. So far, no problems.


      However, attached part is an 200x200 millimeter part. The goal is to create a 1x1 micrometer (=0.001 millimeter) surface. This gives SolidWorks a lot of trouble. The creation of a 1x1um surface using a Point-cloud mesh is impossible. By using different threads on this forum i discussed this problem and found some work-around. By creating a larger surface, and using the scale function, it is possible to make an 1x1um surface. However, solving one problems leads to another regarding this scale. When I put the parts in an assembly, I cannot define a contact point since the 'move component' functions does not work.


      So, long story short; Is it possible to apply the FEA on a larger surface (for example, an 10x10 millimeter surface), while the properties of the material also are scaled? Since the parts are a 1000 times as big, the material has to account for that. And if so, does someone know the material properties?


      I really hope this is possible, since I am really struggling with this..

        • Re: Scaling material within Simulation?
          Jared Conway

          i don't think i understand your input problem.


          what are you starting with? what problem does solidworks have with it?


          regarding scaling the materials, you're asking specifically how E needs to change to account for scaling?


          have you already setup some test examples? should be really simple with say a cantilever beam


          another question for you, are you expecting scaled results or the same results with the adjusted materials?

            • Re: Scaling material within Simulation?
              Mark Kamps

              I'm sorry, let me explain it better.


              I am starting with an PointCloud file (txt), generated from a MATLAB code. This mesh will be generated into a surface. This is where the first problem occurs; whenever I import an mesh surface smaller than 1x1 millimeter, SolidWorks gives an error about the unit setup not being smaller than 1mm. Even if I did change the units to nanometers, it still gives the error. So the work around is to scale the PointCloud file using MATLAB (x1000) and than import it. We can than use the surface generator to create a surface out of the PointCloud. I added the files and parts regarding this problem. (The PointCloud attachment is the smaller one which gives errors, the PointCloudFactorized attachment is the scaled one, which can create a surface).

              This problem I just mentioned, has been treated in this forum as well, link: MySolidWorks  - Official SOLIDWORKS Community


              After that, I create a part with this surface, by 'Boss Extruding' up to the surface. This gives a nice solid surface. However, it is in the wrong scale. (currently; 200x200 millimeter, goal; 1x1 micrometer). I am interested in creating a surface of 1x1 micrometer, to use in FEA analysis together with an simple plate part (this plate is just a 1x1 micrometer surface of minimal thickness).


              Now with a lot of work-arounds it is possible to create the surface, but when using the 'Move Component' feature, it wont work. The goal is to use collision detection to locate the contact point, however the surfaces just intersect each other. Even with 'stop at detection' activated.  When you manually make an mate selecting the touching surfaces, and setting up all the parameters to use the FEA, the analysis will fail. I added the parts of the assembly, as well as the settings. (SW2014) (this is an 10x10 micrometer assembly)


              The solver will give the following warning:

              PCGLSS0204:Negative or zero pivot -7.842117e-009 at row 3463 


              This is all on the small scale system (1x1 or 10x10 micrometer). I already did the analysis without problems on the larger scale 200x200 millimeters. On this scale the solver runs without problems and I get the desired results. However, these results are for the larger scale system, which is not usable. I want to get the results for the 1x1 micrometer system. The way I see it, that is possible by either making a 1x1 micrometer surface, or by adjusting the material settings to account for the scaling of the surface. So to use the 200x200 milimeter system, but using a 'scaled' material. So the results are the same as when a 1x1 micrometer surface was tested.


              I know it is kind of confusing.. My apologies.

                • Re: Scaling material within Simulation?
                  Jared Conway

                  1. what is the goal of the analysis? what will the setup be?

                  2. what is the application that is such a small scale? even if you something that is scale 1 and another at 10, you can figure out how to scale for your application

                  3. why can't you use a more "Standard" application like a cantilever beam to figure out how the scaling needs to be done?

                    • Re: Scaling material within Simulation?
                      Mark Kamps

                      I think you don't fully understand me. The biggest problem isn't the scaling, but the analysis not working on this scale. It gives error it doesn't give on a larger scale system. So it is possible to make a scale 1 system, and to make a scale 10 system (both with their own 'glitchyness' in the software). So the problem isn't the scaling.


                      The problem is that the analysis will fail, and the meshing also gives problems on this small scale, while these problems don't occur on the large scale.


                      So to answer your questions;


                      1: The goal is to extract the forces that are acting on the compressed surface. The bottom part is an micro scale surface (so how an actual surface looks with an very strong microscope). The top part is an flat rigid plate. The top part will be pressed onto the bottom part with FEA. The plate will be pushed onto the surface with an fixed distance. The goal is to extract the resulting/normal forces, as well as the deformed surface.These forces and the displacements in the surface give a unique graph which I need.

                      2: The application is a real life surface, as looked under an microscope. As said before, the scaling itself isn't the biggest problem. However, on this scale, SolidWorks becomes sort of 'glitchy', where clicking on surfaces gives a miss click, it appears wrong, and the meshing and analysis itself wont properly work. All this, will work fine on the bigger scale, since we already did everything there. The goal is now to scale the system down, to get the results.

                      Since the system won't run on this smaller scale, the idea was to perform the analysis on a larger scale, but account this scaling in the material properties. Is there a known way or work-around to account for this scaling in the material properties?

                      3: See above..

                        • Re: Scaling material within Simulation?
                          Jared Conway

                          have you tested on a sample part that you can get the results you want? reaction forces only come out of restraints. based on what you want to do you're going to get contact pressure.


                          regarding the scale, you may be a the resolution of the modeler based on the miss clicks. some good articles about that in the solidworks kb. scaling up is the only option, eventually as you rebuild surfaces will start to fail because of rounding...etc.


                          as for scaling up, I think the best suggestion is to build a couple tests with known answers and see what needs to be done. as I mentioned above, a cantilever beam that is 1x and 10x should tell you what you need to do. then you have an example to go along with your method description.


                          just thinking it through, lets say you had a 1x1 block 10inches long and apply 1lb, lets say you get 1inch of displacement

                          if you have a 10x10in block, still 10inches long, to get 1inch displacement, you're either going to have to make the load way higher or make the material stiffer. is the scale 10x? that's what i'm recommending checking and adding to your references.