Electricity Consumption Forecasting Using Gated-FCN With Ensemble Strategy

Accurate electricity consumption forecasting in the power grids ensures efficient generation and distribution of electricity. Keeping this mind, paper introduces a novel deep learning model, termed Gated-FCN, for short-term load forecasting. The key idea is to introduce an automated feature selection model includes eight-layered Fully Convolutional Network (FCN-8) which hand-crafted that requires expert domain knowledge avoided. Furthermore, also reduces noise as it learns internal dependencies correlation time series. Enhanced Bidirectional Gated Recurrent Unit (EBiGRU) used combination with FCN-8 learn long-term temporal Moreover, weighted averaging mechanism multiple snapshot models adopted proposed assign optimized weights BiGRU. At end BiGRU, fully connected dense layer gives final prediction results. Gated-FCN end-to-end does not require any other enhancing its efficiency. Different activation functions are initially analyzed determine how complex patterns from series data. Later, function having best accuracy extracts both spatial features provides predictive exploratory data analyses assist policymakers making optimal decisions regarding production dispatch. In order demonstrate applicability technique, simulations performed using nine years’ taken Independent System Operators New England (ISO-NE). comparison five state-of-the-art techniques provided prove fact compared benchmark terms two performance metrics: Mean Absolute Percentage Error (MAPE) Root Square (RMSE).