Add method for imputing load-following generation

[grgmiller]

For fossil generation for which we are missing EIA-930 profile data, we are currently assuming a flat profile (essentially using the monthly average value). However, in some cases, it would likely be more accurate to model this missing generation as load following. To do so, we would assume that

For peakers: Merit-order approach based on load (from Cicala pg 54) Assume the generator produces at max nameplate capacity to satisfy peak load, rank hours of each month by demand, then produce at maximum capacity in the highest load hours up to the total reported monthly generation.

For load following: ???

Both of these methods will require using demand data from EIA-930 as an input, which isn't currently used