Dealing with demand and weather uncertainty in electricity system planning
Energy Futures Lab hosts a seminar from Adriaan Hilbers on weather uncertainty in electricity system planning
Abstract
Since the decarbonisation of electricity is expected to require enormous investments, determining the optimal “pathway to sustainability” is essential. Optimal decarbonisation strategies are frequently determined using power system models: computer simulations of the electricity system and some guiding assumptions regarding e.g. future technology costs or demand profiles. These assumptions induce uncertainty in results since other scenarios may occur in reality. One such uncertainty comes from the use of demand and weather data. Recent studies indicate that this effect is significant; for example, estimated optimal decarbonisation strategy may depend heavily on e.g. which year of demand & weather data is used. Hence, picking the “wrong sample” of demand & weather data may lead to a suboptimal energy strategy in the long run. While the obvious remedy is to use longer samples (e.g. longer than a single year), this is frequently unfeasible due to the computational resources required to solve such optimisation problems. In this talk, I will introduce two statistical techniques to mitigate this issue. The first is a subsampling scheme that makes the use of long demand & weather samples in electricity planning possible. The second is a bootstrapping scheme that allows demand & weather uncertainty to be quantified. Combined together, these approaches allow robust decision-making in the decarbonisation of electricity.
Biography
Adriaan is a second-year PhD student in the Statistics section (department of Mathematics) at Imperial College London, as well as the EPSRC Centre for Doctoral Training in Mathematics for Planet Earth. His research focuses on subsampling methods for optimisation models, with particular application to renewable electricity planning. Before this, he studied Mathematics at the University of Oxford and worked for just under a year in strategy consulting in Amsterdam, where he is from. He is a keen science communicator, taking part in speaking competitions, as well as leading the Imperial side of the Energy Journal, a biannual student-led magazine and website created in collaboration with LSE, UCL and ESCP.