Prediction model of electrical energy consumption in conventional residential buildings using ANN and ANFIS

The energy consumption of a residential building is considered in terms of energy use and efficiency. Therefore, forecasting the energy consumption of buildings has been raised as a challenge in recent decades. In a residential home, electricity consumption can have recognizable patterns daily, monthly, or yearly depending on living conditions and daily habits and events. In this research, artificial neural network (ANN) and adaptive fuzzy-neural inference system (ANFIS) have been performed using MATLAB software to predict building energy consumption. Also, random data collected based on the criteria obtained from the hourly electricity consumption of conventional residential buildings in Tehran has been used. In order to evaluate and measure the performance of this model, statistical indicators have been used. According to the applied settings (type of learning, number of steps, and error tolerance), the system error rate is calculated based on MSE, RMSE, μ, σ, and R statistical indicators and the results of energy consumption forecast in three buildings with high accuracy and correlation coefficient. R is more than 98%. The output of this research is an intelligent combined system of ANN and ANFIS. The obtained values well show the ability of this model to estimate energy consumption in the mentioned buildings with high accuracy.