Residential Load Forecasting Using Deep Neural Networks (DNN)

Abstract

Forecasting of consumer electricity usages plays an important role to make total smart grid system more reliable. As the activities of individual residential consumers has many uncertain variables, it is hard to accurately forecast the residential load levels. For planning of the electrical resources and to balance demand and supply, accurate forecasting tasks are critical. This paper presents Deep Neural Network (DNN) based short term load forecasting for Residential consumers. In this work, we compare the Mean Absolute Percentage Error (MAPE) value for residential electricity dataset using different types recurrent neural network (RNN). Our preliminary results indicate that Long short-term memory (LSTM) based RNN performed better compared with simple RNN and gated recurrent unit (GRU) RNN for a single user with 1-minute resolution based on one year of historical data sets.