Stochastic modelling and forecasting of wind capacity utilization with applications to risk management: The Australian case

Wind electricity generation in Australia is accelerating, and South Australia has one of the highest penetration rates globally, so pricing and hedging risks associated with wind power generation are becoming of critical importance. This paper proposes stochastic models for the dynamics and attributes of wind capacity utilization, a typical underlying asset of wind derivatives. Using daily wind generation data from the five states of the Australian national electricity market, we will identify the nature of wind capacity utilization (generation/capacity) in terms of seasonality, mean reversion, and roughness. We employ rough and Lévy stochastic models to capture these dynamics and gauge their efficiency for calibration applications. Finally, we will assess the forecasting performance of the proposed models. This research will offer practical insights into the role of wind derivatives in the energy transition and the integration of wind energy into the electricity markets.