Implement frame-by-frame energy correction for fit-grains

[psavery]

We can keep using the current beam energy in the instrument class as an "average" beam energy. Most parts of HEXRD will work fine with this average beam energy. For example, for the indexing step, it is probably okay to use the average beam energy.

However, for fit grains, we also want to perform frame-by-frame correction based upon the different energies of each frame. We should look into how we might be able to take into account this list of energies for fit-grains, without having to refactor major parts of the code. Perhaps one step would be to add an argument to simulate_rotation_series() with the beam energy list. That might even be enough for all of pull_spots() to work...

See discussion in #592.