openego / eDisGo / 33615567553

84%

dev: 84%

push

github

fix(run): accept eGo's overlying-grid SoC key names and report what is dropped

`import_overlying_grid_data` matched the keys of the passed overlying-grid
data against the attribute names of `OverlyingGrid` and silently ignored
everything else:

    for attr in edisgo.overlying_grid._attributes:
        if attr in overlying_grid_data:
            setattr(edisgo.overlying_grid, attr, overlying_grid_data[attr])

eGo's interface emits `thermal_storage_central_soc` and
`thermal_storage_decentral_soc`, while the attributes here are
`thermal_storage_units_central_soc` / `..._decentral_soc` — the same data, the
name differing by the `units` infix. Both frames were therefore never set, in
every eGo run, and nothing said so. 11 of the 13 attributes arrived; these two
did not.

The consequence is not a missing input but a wrong constraint. With the frame
empty, `_build_heat_storage` falls back to `data = 0.0` for the whole state of
charge, and `soc_initial` and `soc_end` of every thermal storage are taken
from the first and last row of it. So each storage is forced to start empty
and end empty instead of following the schedule eTraGo planned. Within a
horizon the storage can still shift heat, but it cannot carry any across the
boundary — which bites hardest under temporal complexity reduction, where the
horizons are short and the boundary dominates. Thermal storage is one of the
flexibilities the OPF exists to dispatch.

Rather than renaming in eGo, the alias is resolved here, for three reasons:
the silent drop is on this side and has to be fixed here anyway; the CSV path
reads files named after the attributes, so exports already written under the
eGo names stay unreadable if only eGo changes; and the two release cycles are
independent. `ATTRIBUTE_ALIASES` in `network/overlying_grid.py` maps the
deprecated names, and both the runner task and `OverlyingGrid.from_csv` accept
them. Every use logs a warning naming the attribute to rename to, so this
stays a migrati... (continued)

24 of 25 new or added lines in 2 files covered. (96.0%)

9479 of 11318 relevant lines covered (83.75%)

0.84 hits per line