Combinatorial Data‐Driven Innovation of Ecofriendly Transparent Solar Heat Control Coating for Green Buildings

ABSTRACT Transparent solar heat control (TSHC) coatings for windows have garnered significant attention as a key technology for passive cooling in green buildings to reduce energy consumption. However, many studies have focused only on TSHC coatings composed of single functional nanoparticles, and the development of these coatings traditionally relied on trial‐and‐error methods. Herein, we propose a real experimental data‐driven tandem neural networks (NNs) model, comprising spectrum NNs and inverse design NNs, for the combinatorial innovation, development, and optimization of TSHC coatings. Attributed to the high quality of the data, the resulting well‐trained tandem NNs with an R² value above 0.95 facilitate the rapid development and precise inverse design of TSHC coatings with multiple functional nanoparticles. The developed coating, composed of cesium tungsten oxide (CWO), antimony tin oxide (ATO), and indium tin oxide (ITO) nanoparticles, achieves a luminous transmittance of 69%, UV transmittance of 0.1%, and NIR transmittance of 4%. The calculated solar heat gain coefficient (SHGC) and light‐to‐solar gain (LSG) ratio are 0.49 and 1.41, respectively. Temperature reduction tests using a house simulant revealed that the developed TSHC coating can reduce indoor temperatures by up to 8°C. Furthermore, innovative application methods, including spray coating and solution‐processed film techniques, have been explored to apply the TSHC coating to large glass surfaces. Our work provides a novel strategy to efficiently develop and optimize the optical properties of coatings with multiple functional compositions.