Ctrl+K

Rule solve_network

Rule solve_network#

digraph snakemake_dag { graph [bgcolor=white, margin=0, size="3,3" ]; node [fontname=sans, fontsize=10, penwidth=2, shape=box, style=rounded ]; edge [color=grey, penwidth=2 ]; 0 [color="0.64 0.6 0.85", fillcolor=gray, label=solve_network, style=filled]; 1 [color="0.33 0.6 0.85", label=prepare_network]; 1 -> 0; }

Solves linear optimal power flow for a network iteratively while updating reactances.

Relevant Settings#

solving:
    tmpdir:
    options:
        formulation:
        clip_p_max_pu:
        load_shedding:
        noisy_costs:
        nhours:
        min_iterations:
        max_iterations:
        skip_iterations:
        track_iterations:
    solver:
        name:

See also

Documentation of the configuration file config.yaml at electricity, solving, plotting

Inputs#

Outputs#

  • results/networks/elec_s{simpl}_{clusters}_ec_l{ll}_{opts}.nc: Solved PyPSA network including optimisation results

Description#

Total annual system costs are minimised with PyPSA. The full formulation of the linear optimal power flow (plus investment planning) is provided in the documentation of PyPSA. The optimization is based on the pyomo=False setting in the network.lopf() and pypsa.linopf.ilopf() function. Additionally, some extra constraints specified in prepare_network are added.

Solving the network in multiple iterations is motivated through the dependence of transmission line capacities and impedances on values of corresponding flows. As lines are expanded their electrical parameters change, which renders the optimisation bilinear even if the power flow equations are linearized. To retain the computational advantage of continuous linear programming, a sequential linear programming technique is used, where in between iterations the line impedances are updated. Details (and errors made through this heuristic) are discussed in the paper

Warning

Capital costs of existing network components are not included in the objective function, since for the optimisation problem they are just a constant term (no influence on optimal result).

Therefore, these capital costs are not included in network.objective!

If you want to calculate the full total annual system costs add these to the objective value.

Tip

The rule solve_all_networks runs for all scenario s in the configuration file the rule solve_network.

Contents