This paper addresses the problem of short-term load forecasting on an electric power grid. Accurate load prediction plays an important role on multiple aspects of electric grid operation and management, including risk assessment, maintenance and outage planning, coordination between different grid operators, contributing to improve efficiency and resiliency. A prediction model based on artificial neural networks are employed to process data from the Portuguese power grid with a 15 minute sampling interval. In addition to the grid load data, additional inputs were added, including weather information and results from clustering of the time series. The system produces a 24 hour load forecast, for each 15 minute.

--

Electrical grid,Neuronal Networks,Short-term load forecasting