Improve quality of build_hydro

[euronion]

Plotting inflow from build_hydro reveals unrealistic peaks for MD. A similar peak is also present for DE, other countries were not investigated. The peaks in both cases exceed what is possible with existing hydro capacities (DE: factor 2 ca. 4.5 GW installed, MD: factor 5-10, ca. 50 MW installed)

Since these are only temporary peaks they probably don't affect the results too much, also because they are relatively low.

Ideas how to handle:

• Change how the inflow is normalised, e.g. using in addition to the annual generation also the maximum installed capacity per country
• Improve the quality of the inflow dataset
• ?

[fneum]

xref #168