Optimizing transition temperatures for thermochromic smart windows via neural networks coupling simulation

With the global emphasis on reducing energy consumption and greenhouse gas emissions growing, the design and operation of energy-efficient buildings have become increasingly important. Thermochromic has offered a solution to this problem by intelligently adjusting indoor solar radiation and the optical properties of windows in response to real-time temperature changes. However, the optimal transition temperature setting in different climatic zones is still controversial. This study aims to resolve the uncertainty surrounding optimal transition temperatures by coupling EnergyPlus simulations with a BP (backpropagation) neural network model. The application of thermochromic windows in residential buildings across 34 cities in China was investigated to optimize energy efficiency and transition temperatures. By analyzing the impact of varying transition temperatures and basic building parameters across different climatic regions, this study identifies optimal strategies for maximizing energy savings. Building on this, training the neural network model achieves a direct parameter relationship and a faster calculation process. Climate parameters from sources like the National Meteorological Data Network can be input to quickly determine energy consumption and optimal transition temperatures, even in regions lacking complete weather data. This provides a practical tool for engineers to assess the energy-saving potential and optimize the design of thermochromic windows across diverse regions, including areas beyond China.