Structural Health Monitoring and Intelligent Maintenance Algorithm of Building Engineering Based on Graph Neural Networks and Reinforcement Learning

With the acceleration of urbanization, the safety and maintenance of building structures are becoming increasingly prominent, while the traditional structural health monitoring methods are limited by the data processing ability and the intelligence of maintenance decision. Therefore, a new algorithm combining GNNs (Graph Neural Networks) and RL (Reinforcement Learning) technology is proposed in this study, in order to realize efficient and intelligent monitoring and maintenance of building engineering structures. In this article, firstly, a learning model of architectural structure representation based on GNNs is constructed to effectively capture the complex relationship between structural nodes. Then, combined with RL algorithm, an intelligent maintenance decision-making method is designed, which can adaptively select the optimal maintenance strategy according to the structural health state. The experimental results show that compared with the traditional structural health monitoring methods, the algorithm combining GNNs and RL proposed in this study has significantly improved the accuracy of anomaly detection, reaching more than 97%. Furthermore, it shows advantages in maintenance cost and control of building structural risks. The algorithm can not only effectively identify structural anomalies, but also dynamically adjust the maintenance strategy according to realtime data, thus realizing intelligent management and maintenance of building engineering structures.