Electrical Load Forecasting Using Customers Clustering and Smart Meters in Internet of Things

摘要

Forecasting the electrical load demand is one of the main concerns of governments, electric utilities and electrical engineers in industry. While the most of forecasting methods use the aggregated load of system to forecast the electrical load of future, the proposed method in this paper clusters the costumers based on two groups of features: 1- features obtained from the electrical load curve, 2- features acquired from filled questionnaire. The half-hour electrical load variables obtained from the smart meters through the Internet of things technology together with the lagged load and calendar variables are used as input of a multilayer perceptron neural networks to provide a short-term load forecasting. The experimental results show the good performance of the proposed method compared to its competitors.