I am trying to simulate a situation where a big piece of blade first rotates for about half a circle around the engine axis and then is released at a certain point. Its similar to blade-out. Need to calculate trajectory of the piece. I am doing proof of concept so dealing with coarse mesh of around 1.6M mesh cells as of now. I have managed to get the rotation part right, using the 1DOF rotation restriction about specified axis. Works as expected. But as soon as I switch to the release motion where I make the rotation velocity zero and remove 1DOF restriction, keep the linear velocities, specify inertial tensor about the current CG location, the simulation diverges in 2 time steps. I see that its the angular velocities which always run into millions of rad/sec and make the run diverge with excessively high temperature, viscocity ratio and rasiduals. I have tried everything. Reduced time step as low as 1e-10, switch on Solution Stabilisation and Implit mesh update for every iteration etc. Every single time it diverges with same error - runaway anglular velocities in second time step. The only thing that stops the run from diverging is to increase the Moment of Inertia values by 5-10x. Then the angular velocities remain low. I have checked and double checked the Inertia tensor values, they are correct (taken from LS-Dyna). If I run the simulation for say 100 time steps with 10x the actual MOI, and then change the MOI values to the actual, the run again diverges in 2-3 time steps.
I am not able to figure out what is wrong with the simulation. I am using Pressure based coupled solver, default solver settings and discritization methods and so on.
My issue is similar to the one in this thread - /forum/forums/topic/moment-of-inertia-issue/
However, switching on the Implicit Update has not helped in my case.
You are navigating away from the AIS Discovery experience