In this repository, we find buffer capacity factor based on our proposed approximate approach c_a (cf. rule compute_approximate_c),
robust approach for each subset c_r (cf. rule compute_robust_c) for each subset and whole of the network,
and line-specific approach c_l (cf. rule compute_linespecific_c) for the all transmission lines.
This repository is useful to test the operational costs of
a network with different buffer capacity (contingency) factors (cf. rule solve_heuristic_contingency) versus the
fully security-constrained case (cf. rule solve_full_contingency), as well as how
secure the heuristic systems for specific outages (cf. rule check_outage_flow).
Documentation inside scripts/*.py.
The configuration is set in config.yaml.
network:specifies the path to the already solved PyPSA-Eur network. Ideal would be a 50-node German network with investments optimized for 90% CO2 redunction, which has lots of flow, but also some conventional generators to set high marginal prices.load_shedding:adds load shedding generators to the network if not already existing to guarantee feasibility.rolling_horizon:runs full security-constraint LOPF in batches of snapshots rather than all at once.group_size:specifies the number of snapshots that form a batch.s_max_pu:specifies which heuristic contingency factors you want to sweep through.solver:includes the solver options and parameters.
Mount the LSDF, do kubectl-login, go to project directory and execute:
kubedo --inception -i pypsa-eur-gurobi -- snakemake -j 99 all
Make sure you do not have any lines with num_parallel=0 and infinite reactance x=np.inf.
Make sure you do not have sub networks with a single bus.
You can get the operational costs of a network with different buffer capacity factor with notebooks/operation_cost.py.
You can check out the line loadings at the different outages with
import pandas as pd
pd.read_csv("results/new/outage_line_loading_heur....csv", index_col=[0,1,2], parse_dates=True)This file has multiple index levels: the first is the snapshot, the second the component type, the third the line index. The columns denote the outage of the line. The column base shows the flows under no outage conditions. You can compare this data to the line capacities in n.lines.s_nom to find any overloadings.
You can find total cases and snapshots where the transmission network is overloaded in at least one line for different buffer capacity factors with notebooks/check_outages.py.
All results has been plotted with notebooks/plot.ipynb.