Application of thermal energy optimization based on improved neural network algorithm in green innovation capability evaluation of manufacturing enterprises

As the global focus on sustainable development continues to increase, manufacturing companies face the dual challenges of energy efficiency and environmental impact in the process of enhancing green innovation capabilities. The purpose of this study is to explore the application of improved neural network algorithm in thermal energy optimization, so as to improve the green innovation ability of manufacturing enterprises and promote their sustainable development. By constructing an improved neural network model and using big data analysis technology, this paper conducts in-depth analysis of thermal energy use in manufacturing enterprises, so as to identify and optimize energy consumption patterns. The study considered a variety of influencing factors, including equipment efficiency, production process and environmental policy, and evaluated the optimization effect through model training and testing. The experimental results show that the improved neural network algorithm can effectively identify thermal energy waste points, and put forward the corresponding optimization measures. After optimization, the energy use efficiency of manufacturing enterprises has been improved, carbon emissions have been significantly reduced, and the comprehensive evaluation score of green innovation ability has been improved, providing effective technical support for promoting the sustainable development of enterprises.