Speed up optimise energy method

[wouterpeere]

Based on issue #306 it was found that for some profiles, the optimise energy method is very slow to converge, however it works. The problem is that, since the implementation of the EERCombined class, it is no longer a priori known what the ground load will be at a certain time. This issue has the goal of reimplementing this method in another way so it can cope with that challenge.

[wouterpeere]

One solution can be to adapt the calculate_temperature_profile method in such a way that it can take another attribute 'end' so that instead of calculating the entire load profile, it stops after 'end' months. This will reduce the computational time, probably, with a factor 2.