Consistency of peak load at various aggregation levels

[gplssm]

In versioned DP dataset v.0.2.10 peak loads in load areas and LV grid districts deviate quite heavily.

Cumulative deviation in kW:

diff_residential      92549.608809
diff_retail          276115.733138
diff_industrial      -77962.741536
diff_agricultural    338403.227488

Model draft data of v0.3.0 shows smaller cumulative deviation

diff_residential      1.226731
diff_retail          -1.759502
diff_industrial      -5.175576
diff_agricultural    51.138129

But maximum deviation for agricultural and residential sector is large to be acceptable

Maximum deviation 'model_draft' in kW:

diff_residential     5.920596e+00
diff_retail          1.110223e-10
diff_industrial      5.551115e-11
diff_agricultural    5.408894e+01

Fix this in ego_dp_lv_consumption_peakload.sql

In addition, Ding0 does not consider load areas with a peak load smaller than 1 kW.
This raises two issues

1. Data is inconsistent to LV grid district peak loads
2. Data of of EHV/HV is inconsistent with MV/LV peak load data

"1." Could be solved in Ding0, but does not solve "2.". I don't have a proposal how to resolve this. Any ideas @ulfmueller @IlkaCu or @lukasol ? From my perspective its not an urgent issue, but it should be resolved at some point...