Automate process of recording settled states of kangaroo

I want to automate the following process
1 - simulate a system with kangaroo until it settles
2 - record some data on the result with a reference to the input parameters
3 - reset kangaroo
4 - change parameters and repeat until a full set combination of the parameters has been run and saved.

I can’t see a way to get data on when kangaroo settled state is reached - A way around this could be to reset by animation count or run the step solver triggering a new animation frame on every loop with 100 000 iterations
Sounds really inefficient

I’m unsure of the sorta process control logic I could use to run this -
Don’t think anemone is going to work for the loop control as I’m not returning the same data which is feeding into the loop.

Do I have to run the kangaroo solver definition inside a hops component and have a timer sequentially trigger it with new parameters?
Is this something the Batch Driver in Meta Hopper can do?

Any suggestions?

There’s an example in this topic that might help you. Uses data recorder and you can animate a slider to get the kangaroo solution for each slider value

Thanks for the head’s up on the data recorder and data dam method @martynjhogg

I’m able to run short snap versions by animating the main slider to both change inputs and trigger kangaroo step solver
but it’s kinda borking most of the time.

I’m trying to add delays in to the timer triggering to allow re processing of the sketch before the animation is triggered - doesn’t seem to be helping
tried slowing the animation with using a firefly 1 way slider on a timed trigger - Kangaroo didn’t get along with that.

Any suggestions on how to slow the animation input control down?

Adding sub-iterations on kangaroo step solver slows the animation - but still doesn’t allow the so canvas to compute before updating kangaroo (Or at least I think this is what’s happening)

Thanks again for the suggestion

I’ve managed to get some useful output now by keeping it simple and animating just a force/goals parameter for the kangaroo input… This way I can at least get a range of handy looking outputs for a setpoint design parameter
and then just iterate running the calculation after I vary the line network configuration manually.
Thanks again

You know more than me now!

1 Like