Short-term electricity price forecasting

Price forecasting has become an important tool in the planning and operation of restructured power systems. This paper develops a new short-term electricity price forecasting scheme based on a state space model of the power market. A Gauss-Markov process is used to represent the stochastic dynamics of the electricity market. Kalman and H_∞ filters, two methods based on the state space model, are applied in order to estimate the electricity price and compare the quality of their state estimates. Our results show that performance measures for the H_∞ filter are generally superior to those for the standard Kalman filter.