Gurara Dam, Nigeria from Sentinel-2 L1C in false colors

Hic sunt leones: modelling an African power pool

Girolamo Ruscelli

Girolamo Ruscelli, an Italian cartographer and writer, in 1561 drawed this map of West Africa. Today, after almost 500 years, we can access from literally everywhere satellite images with an incredibly high-resolution, like the following one produced by NASA/NOAA Suomi-NPP and download from Worldview.

Suomi

Those two pictures can roughly summarise the impressive advancement in the knowledge of our planet but, still today, there are parts of the world with far less observations than others. In the last months I have worked on the modelling of the West African power system (WAPP, West African Power Pool) and even if I didn’t encounter any “Hic sunt leones” the modelling activity was really challenging.

A power system model is a tool that ingests data (power plant characteristics, meteorological conditions, infrastructure information, energy usages, etc.) and provide answers. In our case the answers were “How much the West African power systems are dependant on water resources?” and “What will happen in the near future in terms of CO2 emissions and water consumption?”.

We have written a report available here and we have also shared the input dataset for the Dispa-SET model. The dataset is available on Zenodo.

The most challenging part has been the collection and preparation of the input data for the power system model, described in the section 3 of the report. Data were missing for many important aspects, for example:

  • No general information about outages and availability of power plants
  • Lack of temporal information (time-series) about national electricity demand
  • No information about the management of hydro-power plants and about the inflows and outflows of data

However, we were able to reproduce the main features of the West African power pool and we could give some preliminary insights on what will change in the near-future. Moreover, thanks to the use of the LISFLOOD hydrological model run by our colleagues in Ispra, I could simulate a large set of different weather/hydrological years in order to sample the influence of climate variability on the WAPP energy system.

The results are very interesting and also reproducible, given that both the input dataset and the unit commitment model used are freely available. If you want to know more about it, I invite you to read it and in case to contact me if you need further information.

Senior Data Scientist