Learning Forum

[aaron.m.wells]

Attempting to model a simple representation of a spring and ball style check valve to determine if/how the ball will un-seat from the seal land under vibration loading. Using the approach of transient structural analysis to capture the non-linear contact between the ball and seal seat. Applying a harmonic displacement via a function input to start simple, but may want to apply transient time history ultimately.  All bodies can be assumed rigid or perhaps linearly elastic for the case of the ball to view contact pressures to confirm contact/separation.  

Q1: Unsure how to apply spring preload? If I resolve spring preload in LS1 with Time Integration ON results give reasonable contact pressures at the end of first load step, but contact pressure goes to zero at start of LS2 when start to apply harmonic displacement. Also, what is the best way to apply zero displacement to the base in LS1 and then apply harmonic base excitation displacement via function in LS2?

Q2: Obviously system will be very lightly damped, but should probably add some for stability. When I apply global damping controls I get contact pressure consistent with ball displacement, but if I use longitudinal damping in spring definition, contact pressure is always zero? If all components of the model are essentially rigid bodies and I only care about the displacement of the ball, what is the best approach for damping?

[dlooman]

Can't explain why you are getting zero contact force in certain scenarios, but a typical approach would be two load steps.  In load step 1 the preload is applied with time-integration off.  A displacement boundary condition can be zero in load step 1 and then non-zero in load step 2 with time integration on.  You also mention harmonic input.  Are you doing a Harmonic or a Transient?