DEEP LEARNING-DRIVEN FORECASTING AND OPTIMISATION OF HYBRID RENEWABLE ENERGY SYSTEMS FOR SMART GRIDS
作者
Riyanka Hazra,Sujauddin Mallick,Priyanshu Chatterjee,Rohit sikder,Ayendu Paul
标识
DOI:10.58532/nbennuraisd6
摘要
A deep learning-driven optimisation framework for Hybrid Renewable Energy Systems (HRES) integrated with smart grid technologies. A hybrid Convolutional Neural Network–Long Short-Term Memory (CNN-LSTM) model forecasts solar, wind, and biogas generation with high accuracy (R² ≈ 0.94), while a Rule-Based Energy Management System (RBEMS) implements a Renewables-First dispatch strategy to minimise grid dependency and operational costs. Evaluated over 24-hour and weekly horizons, the proposed framework achieves a 50.4% cost reduction in daily operations (₹9,766 → ₹4,840) and over 62% weekly savings (₹161,629 → ₹61,372). Results confirm that HRES–smart grid integration offers a scalable, cost-effective, and environmentally sustainable pathway for modern energy management.