Short-Term Price Forecasting for Competitive Electricity Market

Short-term price forecasting in competitive electricity markets is critical for consumers and producers in planning their operations and managing their price risk. This paper adopts artificial neural network (ANN) model based on similar days methodology in order to forecast weekly electricity prices in the PJM market. To demonstrate the superiority of the proposed model, extensive analysis is conducted using data from the PJM interconnection. The factors impacting the electricity price forecasting, including time factors, load factors, and historical price factors are discussed. Comparison of forecasting performance of the proposed ANN model with that of forecasts obtained from similar days approach is discussed. The forecasting error is the major concern for forecaster; a lower error indicates a better result. Accumulative error depends on forecasting period (hourly, daily, weekly, monthly, etc.). It will increase for a longer time forecasts. In this paper, the test results obtained by using the proposed ANN provide reliable forecast for weekly price forecasting as the mean absolute percentage error (MAPE) values obtained for the first and last week of February 2006 are 7.66 % and 8.88%, respectively. Similarly, MAPE for the second week of January 2006 is obtained as 12.92%. Forecast mean square error (FMSE) and MAPE results obtained through the simulation show that the proposed ANN model is capable of forecasting locational marginal price (LMP) in the PJM market efficiently.