Shunts in restoration: fail to require restoration

[noahrhodes]

Currently shunts are "shedable" devices, like loads.

Damaged devices are all required to be repaired, but this does not apply to shunts. Shunts are also not in the objective function, so there is no incentive to restore them. As a result, shunts are left disabled, unless required for feasibility.

In particular, for a DC formulation shunts are left disabled even in the final time period.

Should shunts to be added to the objective function?

[ccoffrin]

My initial thoughts.

Shunts that have an impact to active power balance should be added to the objective function for restoring active power.

I think it's best to leave restoration status out of the objective function and leave it a feasibility-only criteria for some components combined with a "restore all" requirement in the last time step.

In the case of the DC formulation we would expect the shunts to always be restored after 100% of load is met, due to a "restore all" constraint but it is ok that it does not understand the feasibility requirements of shunts. That is a fundamental limitation of the model.