openego / eDisGo / 33633976850

84%

dev: 84%

push

github

fix(io): align the thermal-storage SoC before building the heat storage

_build_heat_storage passed the overlying grid's thermal-storage state of
charge straight into a DataFrame whose index is the time index plus the
end-of-period boundary step:

    pd.DataFrame(columns=hps, data=soc.values,
                 index=timeindex.union([timeindex[-1] + timeindex.freq]))

Any series that is not exactly len(timeindex) + 1 long therefore raises

    ValueError: Shape of passed values is (n, k), indices imply (m, k)

That is the normal case for a temporally reduced multi-interval run. Before
the previous commit the container held the whole reduced index (too long);
with the per-interval slicing it holds the interval, and a non-final
interval's trailing boundary step falls in the gap between intervals, so the
slice is one step short. Both shapes crash.

The battery-storage path a few hundred lines above already solved this with
align_series_to_timeindex(..., extra_step=True) plus ffill/bfill for the
boundary step -- the same expression that builds the target index here. Use
it for the decentral and the central thermal SoC too. reindex only touches
rows, so the central lookup by district_heating_id still finds its columns.

The regression test covers both the too-long and the too-short case and
fails on dev with the ValueError above.

5 of 5 new or added lines in 1 file covered. (100.0%)

9478 of 11307 relevant lines covered (83.82%)

0.84 hits per line