Enhancing Solar Irradiance Estimation for Pumped Storage Hydroelectric Power Plants Using Hybrid Deep Learning

Abstract

This research article explores the potential of Pumped Storage Hydroelectric Power Plants across diverse locations, aiming to establish a sustainable electric grid system and reduce per-unit energy costs. A distinctive feature of the study involves forecasting solar irradiance on large-scale hydroelectric dam locations to identify optimal sites for a PV-integrated hydropower system. The research focuses on advancing the integration of floating solar power modules on water storage systems in eight selected regions across India, emphasizing precise solar irradiance estimation. The paper introduces a state-of-the-art hybrid intelligent deep learning model, combining time series analysis and deep learning through residual ensembling to address these challenges. The primary objective is to pinpoint the optimal location for a Power Storage System (PSS) with the highest solar irradiation for PV-integrated hydro system integration. A secondary goal involves minimizing errors within computational time constraints by the proposed model. The study also employs various optimization techniques to enhance its effectiveness and fine-tune the model's performance, contributing to the advancement of sustainable energy solutions. The proposed model performs best with a Whale optimization algorithm with mean absolute error varying from 0.34 to 3.63 W/m 2 and root mean square error from 0.75 to 9.51 W/m 2 on PSS locations. The analysis also confirms average solar irradiance is high on PSS 7 with 221.0 W/ m 2 followed by PSS 1 with 221.1 W/ m 2 among the eight designated sites.