ElectroHydraulicPTO solver stability

[andermi]

Something was causing numerical instability in the ElectroHydraulicPTO "rpm/pressure -> winding current -> torque" solution

To start, I updated the volumetric and mechanical efficiency maps per vendor efficiency maps codified in matlab by @hamilton8415 and brought in a different interpolator: Splinter

@hamilton8415 and I will investigate/resolve issues in this PR

[andermi]

@hamilton8415 something else that was crashing... when the discriminant was negative in the solution for VBus

https://github.com/osrf/buoy_sim/blob/af87b95ea6e4e1a3e67385985cec6d9f1adfb8ed/buoy_gazebo/src/ElectroHydraulicPTO/ElectroHydraulicPTO.cpp#L335-L337

[andermi]

requires osrf/mbari_wec_utils#26

[andermi]

@hamilton8415 can you check what I did with the new interp and see if the plugin produces the correct output?

[hamilton8415]

Will do, I actually wanted to speak with you about how best to incorporate the test comparisons I previously did with the data from the test machine.

[andermi]

@mjcarroll @chapulina @quarkytale the new interpolator (if we decide to keep it) has an MPL-2.0 License, and I was wondering the best way to credit their project?

[chapulina]

IANAL, but reading Q9 here, I think all we need is to include the license file? @tfoote may have a better idea.

[tfoote]

Yes, we should bring in the license copy too.

For clarity I would actually suggest packing splinter as a separate package and then just using it in this package to allow easier tracking of potential upstream changes as well as keeping the licensing differentiation clearer.

[andermi]

splinter as a separate package

I've got it in its own package now.
TODO: I assume I should set the license of the package to MPL-2.0 and put the license file at the top level of the package

[andermi]

requires osrf/mbari_wec_utils#28
and required by osrf/mbari_wec_utils#28
... need to resolve cyclic dependency

edit: moved splinter to buoy_msgs

[andermi]

We should merge this in soon as this has been out quite a while and more branches are starting to use it as a base.

[mjcarroll]

We should merge this in soon as this has been out quite a while and more branches are starting to use it as a base.

Agree, what other things need to get in to call this done?

[andermi]

what other things

I think we are good as is. @hamilton8415 has other branches with some other tweaks coming soon. This PR was a bit nebulous and ill-defined since we didn't know what was wrong exactly. This PR basically turned into "Let's use a different interp"

[quarkytale]

Works for me, CI should come out green as well

[andermi]

@hamilton8415 has requested I cherry-pick new motor efficiency values into this PR before we merge.

[hamilton8415]

Yes, let's merge this as is, then I can issue a PR for the test stuff seperate from that. We will also then probably need to revisit a third PR that fixes an issue or two with the ElectroHydraulic modelling of a battery and current limits, etc...