A diversified Machine Learning based Techno-financial Prediction of a Multi-Energy Hybrid System with the aid of HOMER Simulation Tool.

Abstract

This research focuses on analyzing grid-dependent hybrid energy systems under various configurations to provide reliable power to an office building in Madurai, Tamil Nadu, India.Analyzing insights into a hybrid energy system design with rooftop solar panels, a diesel generator, battery packs, and a converter is the principal target of the investigation. To observe the optimal design, this research considered both technical and economic variables. In this study, cost of energy and renewable fraction data is collected for the considered commercial building resource components using the Homer tool. The regression model was compared with three evaluation matrices of the prediction models: ridge, polynomial, lasso, and decision tree to verify its performanceconcerning techno-economic factors. The findings of the study demonstrate that the decision tree regression model works better than the other three models, with lower mean square error and root mean square error values and significantly greater R-squared values as well.