Optimizing grid connected PV-Battery system for residential energy management system using deep learning

In this research, an IOT based photovoltaic energy management systems (IPOEMS) using deep learning has been proposed for the prediction of the State of Charge (SOC) of the battery. Initially, Temperature, voltage and current parameters are measured using IOT sensors and the data are processed in Arduino Uno for SOC prediction. The proposed method employs a swish-based Temporal convolution neural network(S-TCNN) which predict the power generation and load demand. The Eurygaster optimization algorithm is used to improve the accuracy of the predicted SOC value of the battery. The SOC value is securely transmitted through the grid service provider using AES encryption, which ensures confidentiality and prevents unauthorized access during data exchange. The proposed technique shows that an affordable cost value of 0.9 × Rs and low energy conversion loss value of 1.2% is obtained and compared with other existing methods.