Research On Damage Identification Of Bridge Structure Based On Dynamic Parameter Method And Machine Learning

Bridge is a very important transportation infrastructure,which is conducive to the convenient travel of the people,and also plays a huge role in the rapid economic development of the country.However,the damage of the bridge structure will affect the traffic performance of the bridge.When the structure is damaged seriously,it will even lead to the collapse of the bridge and serious safety accidents.Therefore,it is necessary to accurately identify the bridge structure.Compared with the traditional pattern recognition method,machine learning method has greatly improved the performance of structural damage identification.However,the research on the application of machine learning method in the field of bridge damage diagnosis is still lacking,so the application of machine learning method in bridge damage identification is further studied.Main research work:firstly,the research status of bridge structure is analyzed,and the dynamic parameter method of bridge structure and the commonly used machine learning method are introduced.Then,taking the finite element model of three span continuous beam bridge as an example,the natural frequency of the structure is used as the damage identification quantity of the foundation structure,and the structural damage identification method based on the structural vibration modal analysis technology is established to determine that the natural frequency of the structure can be used as the damage identification quantity.Taking the finite element model of the pre damaged reinforced concrete model as an example,the curvature mode of the structure is used as the damage indicator,and the damage method based on the curvature mode analysis technology is established to determine that the curvature mode of the structure can be used as the damage indicator.Finally,the finite element model of the cable-stayed bridge with low tower is established to simulate the cable damage of the cable-stayed bridge,and the comprehensive identification index of structural damage is constructed.The recognition accuracy of three machine learning methods(BP,RBF,SVM)is compared.In conclusion,the natural frequency and curvature mode of the structure can be used as the damage identification index of the structure,and the three machine learning methods selected can also be effectively applied to the damage identification of the bridge structure,which has a certain application prospect in the field of bridge damage identification,but its extensive application in the damage identification of the bridge structure needs to be further studied.