Forecasting for Temperature Dependent Power Generation Using Kalman Filtering

Power generation forecasting has always been a very important tool for the efficient management, operation and planning of small and large scale power systems. The smart grid architecture featuring electricity markets and distributed generation, with increasing penetration of Renewable Energy Sources, requires more accurate and faster forecasting algorithms. Meteorological parameters affect power generation in the case of conventional as well as Renewable Energy Sources. In this work, we propose the use of Kalman filters to correct the temperature forecast used to determine the expected output of a conventional gas turbine of a power plant participating in the wholesale electricity market in Greece. Time varying, time invariant and steady state Kalman filters are derived. The efficiency of the algorithms is tested through simulations. It is shown that the accuracy of the temperature forecast is significantly improved. The effect of the filtered temperature forecast on the expected of power output of a gas turbine is discussed.