This tutorial provides explanation for the use of the osira codebase.
Firstly, a summary of the main primary parameters, secondary parameters and outputs is given.
Parameters |
Description |
|---|---|
data |
Electricity substation information |
num_substations |
Number of electricity substation nodes |
probabilities |
Cumulative probabilities we wish to use |
iterations |
Number of Monte Carlo draws |
results |
All iterations generated |
data_indirect |
Lookup table for indirect network connections |
Secondary Parameters |
Description |
|---|---|
population |
Direct population disruption |
substation_ids |
Node ids based on the runner index |
substation_asset_ids |
Secondary node ids based on the runner index |
rank |
The iteration rank |
cum_probability |
Cumulative probability |
Railway Station |
Number of stations indirectly affected |
Gas Distribution or Storage |
Number of gas assets indirectly affected |
Outputs |
Description |
|---|---|
output |
Contains all generated results |
The model can be run using either the demo data provided via demo.py or by generating real data via preprocessing.py and using run.py. Each will be discussed in turn.
Synthetic data is generated via the demo.py script and utilizes the given functions within the osira source code.
Once you execute the code below…
python scripts/demo.py
The direct and indirect effects of each event are written to all_results.csv within the results directory.
Additionally, the cumulative probabilities for the specific scenarios are estimated and exported within cp_scenarios.csv.
The results produced are visualized using vis.py in order to plot all event combinations, given their cumulative probability and level of societal disruption.
In contrast to the demo, actual empirical data can be obtained and used to run osira.
However, the data are not released with osira as you probably will require a licence.
For example, the repository has been developed using Ordanance Survey MasterMap data. Should you have access to this data, you can then run the preprocessing script, as follows:
python scripts/preprocessing.py
As each function is run, there will be a message printed to the console explaining the preprocessing steps being undertaken. This involves processing the different types of assets, connecting each asset to an electricity substation and finally exporting the data.
Additionally, local statistical area boundaries are obtained and merged with population estimates. The datasets required are:
Local area boundaries via the ONS geoportal: https://geoportal.statistics.gov.uk/
Population estimates via NOMIS: https://www.nomisweb.co.uk/
After merging the data, it is then possible to use the centroid of each polygon to estimate the proportion of the population served by each electricity substation.