hexagon logo

Using ADAMS/CAR calculated acceleration as input to a NASTRAN model

Hello,
I am looking at extracting accelerations on the body to be used as input to FEA (NASTRAN) body to calculate the stresses. Acceleration is calculated at the chassis.
I am running the model on digital tracks and using F-TIRES.
I am facing the following issues
1) If I consider the chassis to be a Rigid body, the accelerations (at attachment points) calculated die down after 20Hz. Measured acceleration (from testing) still has energy after 20Hz.
2) If I convert my Rigid chassis with a Flex body, I observe spikes in the acceleration vs time graph. These spikes do not go even when I use the SI2 solver or HHT.
3) Flex body does not simulate after reaching a point when General bushing definition -bushes are used. It runs well when the standard bushes are used.
Require help on
1) Any solver settings that I need to try for the above issues I am facing?
2) Whether it is a good idea to use acceleration calculated from ADAMS to be used as input in the FEA model?
3) Any other general help if people have observed such a phenomenon in your simulations and found a way around it?
Thank you
 
Parents
  • 1) Makes sure that your solver has converged, i.e. decrease error until displacements does not change any more. Then decrease Hmax until accelerations does not change any more. GStiff/SI2 should be the most accurate solver for accelerations.
    2) In general, it would be better to use displacements (or even velocities) than accelerations. Adams iterates each step until displacements and forces has converged. But there is no explicit error control on accelerations.
Reply
  • 1) Makes sure that your solver has converged, i.e. decrease error until displacements does not change any more. Then decrease Hmax until accelerations does not change any more. GStiff/SI2 should be the most accurate solver for accelerations.
    2) In general, it would be better to use displacements (or even velocities) than accelerations. Adams iterates each step until displacements and forces has converged. But there is no explicit error control on accelerations.
Children
No Data