Lossy Solitaries

Code repository accompanying the publication [1]: "Network-induced multistability: Lossy coupling and exotic solitary states"

This project contains the python scripts needed reproduce the numerical experiments and respective figures from the paper.

Requirements

The implementation uses pyBAOBAP for MPI-parallelisation. Contact [email protected] for support.

Developed and tested on Python >=3.6. Package requirements are listed in the requirements.txt file. Execute pip install -r requirements.txt in a terminal to install them (though for pyBAOBAP see above link).

How-To

python sampling.py $topology $mode

where topology is one of

• northern for the Scandinacion power grid data set
• circle for the symmetric circle topology
• synthetic for a single realisation from the synthetic power grid model [2].

and mode is

• test for a test run of an experiment to identify bugs
• debug for a debug run with a small number of simulations
• empty for regular simulations

The following simulations are performed for each parameter setup:

• average single node basin stability, e.g. Fig. 2b in [1]
• phase space picture, e.g. Fig. 2a in [1]

Structure

The repository consists of a top-level file run_numerical_experiment.py, that has to be executed to run the numerical experiments, and four subfolders:

• The folder src/ contains source code libraries for the simulation scripts.
• The folder input_data/ contains the network data for the Scandinavian grid.
• The folder simulation_data/ is the target of all data output from a simulation.
• The folder figures/ contains all generated figure files.

src

This folder contains library files that are called from run_numerical_experiment.py.

• launch_parallel_simulation.py: launch a parallel simulation using pyBAOBAP
• parameters_and_fixed_points.py: routines to setup network topologies with different parametrisations, fixed point search
• plotting.py/awesomeplot.py: convenience functions for plotting
• rpgm_neonet.py: generation of synthetic power grid topologies
• sim_functions.py: misc helper functions to organize data output, define numerical observables or perform data clustering

input_data

origin and specifics of data, dynamical regimes

• northern.json: network topology of the Northern power grid [1]
• area-1.txt/area-2.txt: numerical continuation data, contact Roman Levchenko [email protected] for detail
• dynamical_regimes.py: specifications for the parametrizations used in [1]
• northern_dispatch.py: allocation of producers and consumers for the Northern power grid, results are stored in northern_dispatch.csv

simulation_data

This folder is empty. Every time a simulation is performed with run_numerical_experiment.py, the output data is stored here.

figures

The figures can be recreated using the Jupyter notebooks with the respective name. Note that first you need to perform the necessary simulations

Extended Experiments

Run selected parameter regimes

You can run a simulation for each dynamical separately, here it isdynamical_regimes[1] for the Northern power grid, in the following way:

topology = "northern"
dr = dynamical_regimes[1]
single_regime(topology, dr, out_folder="northern_global", number_of_nodes=236, run=["global"])

In this example, a global basin stability evaluation is performed. Pass run=["asbs"] for or run=["pp"] (or any combination) to obtain average single-node basin stability and a phase space portrait, respectively.

References

[1]: "Network-induced multistability through lossy coupling and exotic solitary states", DOI:10.1038/s41467-020-14417-7, arXiv:1811.11518 [nlin.AO]

[2]: "A Random Growth Model for Power Grids and Other Spatially Embedded Infrastructure Networks", arXiv:1602.02562 [physics.soc-ph]