Research on prediction and optimization of energy-saving performance of building integrated photovoltaics based on digital twinning and multi-objective optimization algorithm

This study proposes a framework for predicting and optimizing the energy-saving performance of Building-Integrated Photovoltaics (BIPV) by integrating Digital Twin (DT) technology with an improved NSGA-II multi-objective optimization algorithm. This approach addresses the challenge faced by traditional designs in simultaneously achieving high power generation efficiency, economic viability, and thermal comfort. Firstly, a high-precision DT model is constructed, which integrates the sensor network of the Internet of Things and the thermodynamic simulation of EnergyPlus to realize the dynamic coupling of optical-thermal-electrical multi-physical fields and minute-level state mapping. Subsequently, a mathematical model incorporating decision variables such as PV area, tilt angle, and azimuth angle was established with the optimization objectives of maximizing annual power generation, minimizing total lifecycle cost, and achieving the smallest deviation in the indoor thermal comfort index (PMV). The improved NSGA-II algorithm with adaptive crossover/mutation probability was employed to search for the Pareto optimal solution set. The case verification of a typical five-story office building in Shanghai shows that the error of power generation prediction of DT model is only 3.2%. The equilibrium solution obtained by optimization only reduces the power generation by 4.4%, while reducing the cost by 20.8% and improving the thermal comfort by 7.3%, which is significantly better than the empirical design and single-objective optimization scheme. The research provides a real-time and credible decision support tool for the detailed design of BIPV, and helps to achieve the goal of building carbon neutrality.