frankengen Posted March 14, 2014 Share Posted March 14, 2014 Hi everyone! I have been playing around with the FEM material in H13. And I have been looking at making it follow animation. It works fine with a simple torus squashing and stretching but once I try something more complex it blow up after 4 frames or so. Anyone have any idea how to get around this?Any suggestions would be much appreciated. Thanks, Frank I provided an example scene which explodes. FemTest.hipnc Quote Link to comment Share on other sites More sharing options...
schiho Posted March 14, 2014 Share Posted March 14, 2014 (edited) You can crank up the "Min Substep Rate" to 30+. Even the "Enable Adaptive Substepping" is enabled, when you define this manually, you should get more correct results. If this doesn't help, in the accuracy tab you can set the "Float Precision" to 64. Edited March 14, 2014 by schiho Quote Link to comment Share on other sites More sharing options...
frankengen Posted March 14, 2014 Author Share Posted March 14, 2014 I found a way of making it not blow up, basically I changed the animation, to a slower one and more random and now it works for 500 frames at least. I wish I knew a way around fixing the first one though FemTest.hipnc Quote Link to comment Share on other sites More sharing options...
frankengen Posted March 14, 2014 Author Share Posted March 14, 2014 Thank you so much Quote Link to comment Share on other sites More sharing options...
frankengen Posted March 14, 2014 Author Share Posted March 14, 2014 Mine blows up with min substep rate set to 1000 and enable adaptive substepping either on or of withfloat precision at 64 Quote Link to comment Share on other sites More sharing options...
tricecold Posted March 14, 2014 Share Posted March 14, 2014 mine also blows up, but luckily, they blow up when stress is too much and only some, so I can delete them, but in the manual it says "In some simulations, some tetrahedrons or polygons turn huge after a collision. In this case, try increasing the Min Substep Rate until the problem goes away. Sometimes, you can stabilize a problematic simulation without increasing the minimum substepping level much, by increasing the accuracy instead. For example, you could increase the Max Iterations Linear Solve (to say 10000) and decrease the Tolerance Linear Solve (to say 0.0001)." but this didnt help me in my case Quote Link to comment Share on other sites More sharing options...
Sierra62 Posted March 22, 2014 Share Posted March 22, 2014 With simulations blowing up in FEM, increasing the min substep rate isn't always the best option. When you have fem breaking against another object such as a solid object, increasing substeps on the solver isn't enough. Increasing the global substeps on the Dopnet is the best way to go. In my experience if the impact point happens between frames, the FEM object can get caught in the collision geo and blow up. Another effective way to do this is to up the frame rate to 120fps for the sim and then bring it down to 24. Quote Link to comment Share on other sites More sharing options...
Recommended Posts
Join the conversation
You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.