A hybrid system-econometric model for electricity spot prices: Considering spike sensitivity to forced outage distributions

With the current trend in deregulation, all electricity markets have been subject to volatile electricity prices, typically in peak season. As markets mature, new financial and operational risk management instruments are becoming available. In order to price such instruments, a model for the underlying price process is required. In this paper a hybrid model is described that contains aspects of the power system, as well as the historical time-series of spot prices observed in the market. By including both aspects of the problem, a model with both economic and system-based aspects is created. Its modular design allows easy adaptation to different markets based on generating system reliability, load patterns, and price histories. The spot price histories are manifest in two probability distributions for these prices, whereas the system specifics are included via load and generation models which are calibrated to the climate and system of interest. This model is tested and validated using the Pennsylvania-Jersey-Maryland Market, and is effective is reproducing out-of-sample market behavior. In addition, we conduct an investigation of the sensitivity of the model to various underlying assumptions, such as the shape of the repair time distribution. Preliminary testing of the model on the Alberta Power Market briefly described here were also successful. These investigations provide important insights into some of the driving factors for electricity price spikes.