Electricity generation capacity expansion under uncertainty and risk

Abstract

In this thesis, we formulate a perfectly competitive model that includes electricity generators, retailers, industrial consumers, and an independent system operator as individual agents in a perfectly competitive game. These agents each simultaneously make their operational and contract decisions to maximise their risk-adjusted profit across the potential scenarios, given the other agent's decisions. We show the existence of equilibria for our model, under some relatively nonrestrictive assumptions.We utilise our model to evaluate transmission expansion decisions under agent risk aversion. We find that surprisingly, the new equilibrium with expanded transmission can lead to lower total system welfare. We then show that these inefficiencies are caused by differences in worst case scenarios for generation agents and the system as a whole. We also show how we can rectify this through the use of contracts. The trading of contracts for difference and Arrow-Debreu Securities help align all of the agents' worst-case scenarios with one another and with the system as a whole. This improves the capacity expansion decisions made by the generation agents, making the transmission line a net benefit, even within this case study. We apply our model within the context of a few extensions. We create a case study where generation agents competitively divest from generation. Secondly, we introduce a stage before expansion, where generation agents need to invest in consent and research into their generation plants (or else they cannot build the actual generation plant). Finally, assuming there is a known and proportional impact of carbon emissions, we show how risk-aversion complicates determining the right tax or subsidy when trying to maximise total risk-adjusted welfare. We use this model and aspects of some of our extensions within the framework of the New Zealand Electricity market. Assuming agents are price takers, we model which generation plants we would expect to be constructed or divested from long term. Again, we demonstrate how a liquid contract market impacts investment decisions and how this compares with vertical integration. We compare this all to the socially optimal solution.