Contribution type: Lightning talk & Poster
Modelling VRE forecast errors in large-scale energy system planning studies: Are we planning for enough flexibility means?
Presenter: Etienne Cuisinier
Description: Long-term energy system optimisation models are widely used to derive cost-optimal investment pathways at continental scales; Forecast uncertainty is often simplified in planning studies, including in detailed operational models. As a result, capacity portfolios that appear adequate in multi-weather year assessments may still face short-term flexibility shortfalls when forecast based schedules are exposed to realised VRE generation deviations.
This work introduces an approach using cascading models that propagate variable renewable energy (VRE) forecast errors from day-ahead scheduling to intraday and real time flexibility assessment. Starting from a capacity-expansion framework, the adequacy of the derived capacity portfolio is evaluated with a scheduling layer driven by day-ahead VRE forecasts across multiple weather years; using spatiotemporally correlated forecast errors. The resulting forecast-error-driven imbalances (FEDI) are decomposed into intraday-adjustable and post-gate-closure components, and are evaluated against the residual upward and downward margins available from flexibility means (dispatchable units, transmission lines, VRE curtailment, flexible demand). The proposed framework provides a transparent, tractable stress-testing layer to go beyond usual adequacy concepts when planning for European energy systems with high VRE shares.
The results show that portfolios with low loss-of-load expectation (LOLE) under perfect-foresight multi-weather-year adequacy assessment can still experience non-negligible hours of insufficient flexibility once forecast errors are represented explicitly.
Future work include a more detailed modelling of Balancing and Area & Frequency Control steps; as well as feedback to the capacity expansion model.
Background: This PhD project is hosted at DTU and is part of the MuESSLi project within Cresym; involving 5 PhDs from DTU, TU Delft and UP Comillas. It is funded by RTE, Total Energies and NaTran.
Authors: Mohammadhassan Bahmani, Thomas Heggarty, Etienne Cuisinier, Yalin Huang, Matti Koivisto.
Optional links:
- MuESSLi project: https://cresym.eu/muessli/
- The cascade of models (so called “Balancing Tool Chain”) used in this work relies on the open-source modelling tool Balmorel: Balmorel — Balmorel 5.0 documentation
- The open-source tool CoRES is used to derive spatiotemporally correlated forecast errors for VREs: https://corres.windenergy.dtu.dk/