Issue with PR #167

[NicolaCourtier]

PR #167 moves the following functions required for design optimisation from an example script into _costs.py:

• GravimetricEnergyDensity()
• nominal_capacity()
• cell_mass()

Conceptually the nominal capacity and the cell mass are part of the Model. This becomes clear when considering the definitions for electrochemical models vs. equivalent circuit models. Therefore these two functions should be moved to a base class for electrochemical models. This results in a useful change to the function call from nominal_capacity(x, model) to model.nominal_capacity(x). Also, when changed from a cost-specific function to a model function, we should change the name from "nominal" to "approximate" capacity because its definition can only be an approximation for the nominal capacity.