Levenberg-Marquardt algorithm-based solar PV energy integrated internet of home energy management system

Abstract

With the emergence of smart grids, the home energy management system (HEMS) has immense prospective optimize energy usage and reduce costs in the residential sector. However, the challenges persist in effectively controlling power consumption, reducing energy expenses, enhancing resident comfort, and optimizing coordination of renewable energy sources (RESs). In this study, a Levenberg-Marquardt (LM) algorithm-based solar PV integrated internet of home energy management system (IoHEMS) is developed. The LM algorithm has been chosen as it outperforms the other two artificial intelligence (AI) algorithms: Bayesian regularization (BR) and scaled conjugate gradient (SCG). With the setup of using 70% of data for training, 15 % for validation, and 15 % for testing, the LM algorithm shows the regression of 0.999999, gradient of 7.8e(-5), performance 2.7133e(-9), and the momentum parameter of 1e(-7). When the trained data set converges to the optimal training results, the best validation performance is achieved after 1000 epochs with approximately zero mean squared error (MSE). The proposed system transforms a conventional home into a smart home by effectively managing four household appliances: Air conditioner (AC), water heater (WH), washing machine (WM), and refrigerator (ref.). The proposed model enables accurate switching functions of appliances and efficient grid-to-battery utilization, resulting in reduced peak-hour electricity tariffs. The proposed system incorporates internet of things (IoT) functionality with the HEMS, utilizing smart plug socket (SPS) and wireless sensor network (WSN) nodes. The proposed model also supports Bluetooth low energy (BLE) connectivity for offline operation. A customized android application, 'MQTT dashboard', allows consumers to monitor power usage, room temperature, humidity, moisture and home appliance status every 60 s intervals.