This project aims to improve the accuracy of long-term electricity demand forecasting/demand projections in South Africa. It will achieve this by developing sub-national bottom-up end-use projection models that are sensitive to demand drivers in each region. The model will be spatially disaggregated along the 10 Eskom defined supply areas. The model will also be disaggregated by sector and where applicable by end-use. The structure and calibration of the model will build on existing competencies and models in the Energy Systems Research Group (ESRG) at UCT.
As part of this ongoing task order SANEDI commissioned the ESRG at UCT to develop new long-term electricity demand forecasting models to aid the current and future capacity expansion planning processes, referred to in law as integrated resources planning (IRP). These national demand projections were then further spatially disaggregated using measured Eskom substation data for the 10 supply areas, spatial proxy data, and profile shape correction techniques. The results of these efforts have already been directly used to provide the electricity demand projection scenarios used in the latest national Draft IRP 2024 and 2025.
The Long-term forecasting in South Africa that has been used for supply side development has typically relied on econometric analysis at the national scale. In an era where the energy transition is gathering momentum, it is increasingly important to understand the regional opportunities and constraints of the supply system as well as the impact of a fundamentally changing composition of demand and potential for rapid electrification of end-uses as seen by the grid (e.g. EVs, hydrogen production, self-generation, fuel-switching, increasing cooling demand, etc.). In this context forecasts that rely on econometric methods alone are no longer suitable for long-term scenario-based holistically modelled and optimized energy system planning exercises.
The project will develop hourly electricity demand projections at multiple-scales for various scenario-matched futures, including different assumptions and projections, both for overall growth, technological trends and drivers, and the various dynamic energy demand switching opportunities across all relevant sectors. These will be stored in a database and be publicly available. The energy forecasts developed can be used by Electricity Supply Models such as SATIM, Plexos and PyPSA. Beneficiaries are likely to lie throughout the energy systems modelling community in both the private and public sector, such as Eskom, DMRE, the CSIR, think tanks and generation project developers.