A machine learning based power load prediction system for smart grid energy management

In recent years, the electricity crisis has intensified due to the more electric applications, electrical vehicles, and automotive industries that have been developed. Due to this prediction, power generation, utilisation, and facilities in individual houses, areas, districts, states, and countries are very challenging tasks in a renewable-based smart grid. In conventional methods, they have predicted power utilisation and consumption using advanced metering infrastructure (AMI), big data (BD), artificial neural networks (ANN), and the Internet of Things (IoT), but using these methods has been inaccurate. To overcome these drawbacks, a proposed power management system for the smart grid uses machine learning (PMS-ML) to efficiently determine power usage, generation, and faults as compared to conventional methods. As per the simulation results, the proposed PMS-ML shows better performance as compared to conventional methods with respect to prediction, RMSE, and MSE of 0.25%, 013%, and 0.45%, respectively.