The Study Of Damage Identification On Gravity Dam Based On RBF Neural Networks

China has rich hydropower resources, and China is a big country of hydraulic and hydropower. The construction of the dam has great significance for flood control and drought-resistant socio-economic development. Just like everything in the world, the dam also has life. In general, the dam can be safe for a hundred years, so the dam may face the problem of aging during the run-time. Especially in face of the earthquake, the dam may occur earthquake damage, and the damage may affect its safe operation. If we do not find these diseases, and leave them unchecked. What is worse, dam damage will affect the normal operation of hydropower stations, dam collapse will lead to serious injury factory destroyed, and it will turn water conservation into water damage. Ensure ensuring the normal operation of dams for the protection of lives and property of the people downstream plays a key role. Therefore, they are safety monitoring and damage analysis is particularly important.Damage identification technology is a comprehensive application of science, which has developed rapidly in recent years. Related research at home and abroad is very active, and it is a research hotspot in the field of structural engineering. Structural damage identification is the foundation and core structural health monitoring. Its vibration modal analysis and neural network technology overcome some of the drawbacks of traditional security monitoring methods, which is an effective tool to solve this problem. In this paper, the theory of structural damage identification based on radial basis neural network and its application in the identification of gravity damage did some research, and finally got useful results. This paper describes the principles of the neural network for structural damage identification, and it describes the use of RBF neural network based on the vibration characteristics of structure damage identification methods and procedures. Besides it summarizes the existing research and proposes a two-step damage detection methods of large-scale hydraulic structures. Firstly a method of numerical analysis is chosen. Secondly, the combination parameter is selected as input parameters which is easy to get and which is high recognition accuracy and high sensitivity. The method has a certain accuracy which is combined with shaking table test. The results show that it is feasible to use RBF neural network damage identification methods and procedures to do gravity damage identification.