Enhancing the performance of zero energy buildings with boosted coyote optimization and elman neural networks

Abstract

This research paper aims to optimize electrical system modeling to minimize CO2 emissions, improve reliability, and reduce costs. It proposes integrating Elman Neural Network (ENN) and metaheuristic algorithm to evaluate environmental impact and optimize system performance. The goal is to create a mathematical model considering renewable energy sources, grid energy, CO2 emissions, and maintenance costs, and use metaheuristic algorithms to determine the optimal combination of wind turbines and photovoltaic panels. To train the ENN model and optimize the system, a three-stage approach is proposed. Quasi-random numbers are generated using Sobol techniques to create input data, which is then processed by TRNSYS for simulation. The proposed approach incorporates an ENN with Boosted version of the Coyote optimization algorithm to accurately model and predict the behavior of Hybrid Renewable Energy Systems for Zero Energy Buildings. The results demonstrate the effectiveness of the proposed approach. Our method significantly reduces energy consumption, costs, and system reliability, achieving a 25% reduction in energy consumption and 12% reduction in system cost, and is effective in optimizing hybrid solar/wind and hydrogen storage systems. The optimal system configuration significantly reduces CO2 emissions, achieves a high level of system reliability, and minimizes costs. Comparative analysis with previous studies showcases the superiority of the proposed method in achieving comprehensive optimization. In conclusion, this study combines AI techniques and metaheuristic algorithms to optimize electrical system design, considering environmental, technical, and economic factors. The proposed approach overcomes limitations in previous studies and offers an effective solution to address the optimization problem.