Performance of electricity markets & power plant investments in the transition to a low-carbon power system

© 2018 Dr. Daniel MarshmanIn power systems around the world, substantial quantities of Variable Renewable Generation (VRG) sources such as wind and solar are being introduced. Challenges arise from increasing VRG as it is characterised by significant variability, uncertainty and very low short-run costs of production. This thesis investigates the impacts of increasing VRG within electricity markets at the operational and investment time-scales, particularly with respect to resource adequacy and revenue sufficiency. An electricity market model which uses least-cost optimization is first developed. This contains sub-models covering the generation expansion, unit commitment, and economic dispatch time-frames. The model includes a stochastic representation of forecast errors for VRG and demand. It also includes constraints on Frequency Control Ancillary Services (FCAS) and inertia. The model produces prices for energy and reserves and FCAS, allowing the assessment of the financial performance of generators. The model is also extended to include sub-hourly unit commitment and dispatch. The thesis then examines the performance of generators in an energy-only market with increasing VRG driven by different policies. In particular, the impact of unit commitment constraints on revenue sufficiency for thermal, renewable and storage technologies is assessed. It is found that the presence of wind up to approximately 40% of annual energy does not cause significant deviation from the theory of revenue sufficiency via short-run marginal pricing [131,136], provided the fleet is optimal, i.e. all plant has the opportunity to enter or exit the market. However, fleets with solar greater than 20% of annual energy experience declining financial performance for both solar and thermal generators. The impact of a Rate of Change of Frequency constraint on generator financial performance is also considered. It is found that a 1Hz/s constraint only affects fleets with significant storage capacity, as it is assumed that this technology can contribute to FCAS requirements, but not inertia requirements. A tighter constraint of 0.5Hz/s is also found to impact other fleets which have no storage capacity but significant VRG capacity. How- ever, these fleets already experience declining financial performance for generators as described above. Finally, the impact of the variability of VRG at sub-hourly time-scales on unit commitment and dispatch decisions is investigated. However, this is not found to cause significant cost increases for fleets with either wind or solar up to 60