wuyou33 / bilevel-atc

Cost-optimal ATCs in Zonal Networks

This repository is a companion to the paper [1], and optimizes the Available Transfer Capacity between zones in a zonal network. The scripts are meant to be used with the RE-Europe dataset [2] and scenarios generated as in [3]

Packages needed (Tested version):

- Python 2.7
- Pandas (0.18.1)
- Numpy (1.11.0)
- GurobiPy (6.5.0)
- NetworkX (1.11)
- Matplotlib (1.5.1)
- Seaborn (0.5.1)
- tqdm (4.8.4)

This software was developed on Ubuntu 16.04, and some syntax may be incompatible with other operating systems.

File overview

benders_driver_24bus_3Z.py: Generates plots for the 24 bus test system (Figures 2 and 3).

optimal_driver_stoch.py: Optimizes ATCs for the European test system. Only runs for the hour set in the loadhour variable. To generate the plots in the paper, it is necessary to run this file for loadhour in [1,...,24].

plot_optimized_ATC_layouts.py: Plots the optimized ATC layout (Figure 5) for the set of parameters indicated. Requires optimal_driver_stoch.py to have been run for the same set of parameters.

heuristic_driver_stoch.py: Uses the optimized ATCs, and explicitly evaluates them against all 100 scenarios, along with the heuristics (Static ATC, nodal deterministic). Requires optimal_driver_stoch.py to have been run for the same set of parameters.

optimal_driver_stoch_nodal_direct.py: Generates the optimal stochastic schedule.

save_nodal_stochastic_results.py: Evaluates the optimal stochastic schedule against all scenarios. Requires optimal_driver_stoch_nodal_direct.py to have been run.

plot_functions.py: Plots Figure 4. Requires heuristic_driver_stoch.py and to have been run forloadhour in [1,...,24], and requires save_nodal_stochastic_results.py to have been run.

Solution_strategy.html: Overview of the oslution strategy used to solve large-scale problem instances, as implemented in benders_driver_stoch.py.

Files not provided

Certain files are not provided due to space limitations. These are:

EU_data_scenarios/scenariostore.h5: HDF data store containing signal for load, and scenarios and signals for wind and solar production. The file is organized as

class 'pandas.io.pytables.HDFStore'>
File path: EU_data_scenarios/scenariostore.h5
/load/obs                   frame        (shape->[24,1494])
/solar/obs                  frame        (shape->[24,1494])
/solar/pfcs                 frame        (shape->[24,1494])
/solar/scenarios            wide         (shape->[100,24,1494])
/wind/obs                   frame        (shape->[24,1494])
/wind/pfcs                  frame        (shape->[24,1494])
/wind/scenarios             wide         (shape->[100,24,1494])

With the contents of /wind/scenarios being

<class 'pandas.core.panel.Panel'>
Dimensions: 100 (items) x 24 (major_axis) x 1494 (minor_axis)
Items axis: s0 to s99
Major_axis axis: 2013-06-24 00:00:00 to 2013-06-24 23:00:00
Minor_axis axis: 1 to 1514

To generate these scenarios, see [3]

[1] Jensen, Kazempour, Pinson: "Cost-optimal ATCs in zonal networks" (2016) To be submitted to IEEE Transactions on Power Systems

[2] [https://zenodo.org/record/35177]

[3] [https://github.com/TueVJ/RE-Europe_SimpleEnsembles]