Equilibration times and sampling

[lilyminium]

From https://github.com/openforcefield/openff-evaluator/issues/587 and existing Slack discussion -- this issue is to discuss issues with sampling and equilibration times. The current Evaluator defaults (200 ps equilibration, 2 ns production) may not work well for some of the properties we fit to, such as triethanolamine + water which appears to equilibrate very slowly. The ConditionalGroup protocol is intended to iterate until a target uncertainty is reached, but this isn't sufficient to be catching systems such as triethanolamine.

cc @mrshirts @barmoral

[mrshirts]

Note there are two possible reasons for the slow equilibration - a very slow-moving liquid, and a failure in the box build making the initial volume too high. @barmoral is checking these. Regardless, we may need some method for checking. Enthalpy equilibration may also slightly lag volume equilibration.

[mattwthompson]

Can this be closed now that there is a pre-equilibrated layer?