Forecasting Locational Marginal Prices in Electricity Markets by Using Artificial Neural Networks

Autor: Rosano, Kim Jay R., Nerves, Allan C.
Jazyk: angličtina
Rok vydání: 2021
Předmět:
DOI: 10.22004/ag.econ.333535
Popis: Electricity price forecasting is an important tool used by market players in decision-making and strategizing their participation in the electricity market. In most studies, market-clearing price is forecasted as it gives an aggregated overview of system price. However, locational marginal price (LMP) gives better outlook of the price particular to the customer location in the electrical power grid. This study utilizes Artificial Neural Networks to forecast weekday LMP of generator and load nodes. Various inputs such as historical prices and demand, and temporal indices were used. Using data for selected nodes of the Philippine Wholesale Electricity Spot Market, forecast Mean Average Percentage Error (MAPE) of 6.8% to 6.9% were obtained for generator and load node forecasts, with better prediction intervals than ARIMA models. The results showed that the proposed method of using the LMP of adjacent generator nodes in forecasting load node LMP results in significant improvement of forecast accuracy.
Databáze: OpenAIRE