Pattern Recognition approach to Predict Renewable Energy Consumption

Abstract

The microgrid has been presented in recent years to meet energy requirements as a complementary option. The micro grid comprises of renewable generation, energy storage units and demand management via the low-voltage distribution network, which are part of the intelligent grid implementation. Renewable energy resources, including solar and wind energy, are now used worldwide because of the quick technological progress and environmental benefits. However, it is important to integrate renewable production in the micro grid in advance, for this reason, future power generation from renewable sources must now be predicted. Predicting future energy generation will be helpful in finding requirement for grid integration. Forecasting is the ability to determine periods of stable generation from renewable sources. In this paper a renewable energy generation is predicted which is comprised of solar and wind energy to know the requirements for energy in future. With the help of these results Grid will be configured well in advance to fulfill electricity generation requirements from renewable energy resources. ANN (Artificial Neural Network) and ARMA (Auto-Regressive Moving Average) models are used for prediction of different energy resources. Raw data is first processed using feature extraction technique and then it is used in ANN and ARMA modelling.