RW Jitter modeling validation (N reactionWheelStateEffectors vs 1 reactionWheelStateEffectors containing N wheels)

[bbercoviciUspace]

Hello BSK devs,

We've refactored our simulations such that reaction wheels are no longer bundled up in a single instance of reactionWheelStateEffector but have instead their own instance each.

However, setting the individual RW jitter models to messaging.JitterFullyCoupled while keeping U_s = 0, U_d = 0 doesn't result into dynamics equivalent to the individual jitter models being set to messaging.BalancedWheels .

I would expect the dynamics to be identical, since the imbalance parameters are zero.

Is this interpretation correct ?
Or is RW jitter correctly accounted for only if all RWs are owned by a unique instance of reactionWheelStateEffector ?

[bbercoviciUspace]

The plot thickens : the dynamics are equivalent between the case where the jitter models are set to messaging.JitterSimple while keeping U_s = 0, U_d = 0 and the case where there are all set to messaging.BalancedWheels .

[schaubh]

Mm, that is curious. You are using a series of RW effectors without the RW cluster? That should in theory give the same results. I've never run that like that.

[joaogvcarneiro]

Hey @bbercoviciUspace, I think this is due to the fact that when the reaction wheels are balanced, the effector has no explicit contributions to the system's mass and inertia. See #596 for more info.