Introduce variable for storage losses

[p-snft]

I think it makes a lot of sense, e.g. for adding custom constraints, to have an explicit variable for the losses of the storage. Note that the formula for the relative lasses is more complicated

losses(t) = energy(t) * (1-(1-b)**tau))
energy(t+1)= energy(t)-losses(t)

vs

energy(t+1) = energy(t) * (1-b)**tau

but I think the additional variable and constraint is useful and the 1-(1-b)**tau instead of (1-b)**tau changes a value in the lp file, so it will not significantly increase computational time.

PS: I existentially pushed this change to dev, so this PR reverts the revert.

[Maxhi77]

The possibility to add custom constraint for the losses looks advantageous to me