Consolidate approach to models with RHS time series.

[jd-lara]

In the current code base we have mixed methods on how time series are added to a model to constraint on the right hand side for ThermalStandard (derating), Interfaces, Interchanges, etc. Loads have yet another approach.

After a refactor for DLRs, the approach should be consolidated as implemented here

Sienna-Platform/PowerSimulations.jl#1553

Roughly the approach is

• Add a time series parameter if the attribute is added to the device model
• when adding the Ub/lb constraints or the bounds then call a method to add the constraints with floats or with parameters.

This will avoid unnecessary duplications of constraints.

This is a large undertaking so we will split the work between @luke-kiernan and me.

[luke-kiernan]

Clarifications:

1. Does this mean we're removing ActivePowerVariableTimeSeriesLimitsConstraint and instead just collapsing the keys to ActivePowerVariableLimitsConstraint?
2. For commitment formulations, we currently have two separate upper bounds p <= pmax * forecast and p <= pmax * u. Do you mean that we should combine them into one upper bound, p <= pmax * forecast * u? This will require new functions in IOM.

Currently, upper and lower bounds for semicontinuous are handled together in IOM. However, the presence of a time series only affect the upper bound, not the lower bound. If the answer to (2) is yes, I'll also need to go decouple those.