Fix y intercept of OffsetConverter

[p-snft]

Fixes the OffsetConverter to work as documented.

Closes #1010

[lensum]

Thank you for the quick fix! I haven't looked into the .lp files in detail, but the changes you've made make sense to me.

[Bachibouzouk]

Hello, I actually disagree that this produces the right behavior, I had problems reproducing the graphs of #895 and I stumbled upon this PR.

In the documentation is it stated that P_in(p,t) = P_out(p,t) * C_1(t) + C_0(t) * P_max(p) * Y(t) and it is explicitly stated that P_max(p) * Y(t) is equivalent to status_nominal[n, o, p, t]

I noticed that the beginning of the _relation_rule loop was wrong (

oemof-solph/src/oemof/solph/components/_offset_converter.py

Line 233 in 2edeb22

expr += -m.flow[n, o, p, t]

)

It reads

expr = 0
expr += -m.flow[n, o, p, t]  # this is P_out(p,t) and it is not the one multiplied by C_1(t) !
expr += m.flow[i, n, p, t] * n.coefficients[1][t]

instead of

expr = 0
expr += -m.flow[i, n, p, t]
expr += m.flow[n, o, p, t] * n.coefficients[1][t]

I did locally reversed the changes of this PR (replacing status by nominal_status) and implemented the switch of P_in and P_out and I could reproduce the figures of #895 successfully, which was not the case with the code of this PR.

I need to look at it again with a fresh brain, after considering the points raised in #1010 as well