2002 · 17 citations · 9 references
ComputingComputer EngineeringForecastingMachine LearningReal-time SystemsLoad ForecastingReal-time ForecastingReal-time SchedulingTime Series ForecastingTime Series PredictionDay-ahead ForecastingElectric Load ForecastingHour-ahead Load ForecastingNeural ForecastingOptimal DispatchOptimal Power DispatchPower System ModellingTransactions
The application of artificial neural networks (ANNs) to the real-time load forecasting (RTLF) problem is presented. The term RTLF is used for the prediction of the power system load over an interval ranging from one hour to several hours. This issue is becoming increasingly important with the approach of the open access market with the scheduling of buy/sell transactions as short as half an hour in advance. Separate ANNs are utilized for load forecasting of one hour to four hours ahead. The load forecast of these networks are compared with the of one day ahead load forecast results. Based on simulation results, by utilizing ANN, two objectives are obtained: (1) a more accurate hourly load is predicted; and (2) any near-term buy/sell transactions are fitted in the optimal MW dispatch scheduling. The authors' approach is demonstrated by detailed study of New Brunswick Power data.
9
Short-term load forecasting with local ANN predictors
I. Drezga, Saifur Rahman · IEEE Transactions on Power Systems · 1999 · 136 citations
Computing, Data Science, Forecasting +10
Dipti Srinivasan, Swee Sien Tan, Ching-Chi Cheng et al. · IEEE Transactions on Power Systems · 1999 · 131 citations