The Study On Bridge DaMage Recognition Based On Curvature Modal Theory And Support Vector Machine

As an important part of the transportation infrastructure,bridge is an important guarantee for the rapid development of the national economy.Once the bridge structure is damaged,it will greatly affect the people’s property safety in serious cases.Therefore,bridge structure damage detection has always been a hot spot of research.The beam bridge is the most widely used type of bridge,and it has universal adaptability to its damage research.Support vector machine(S VM)is one of the machine learning methods.It has stronger recognition ability in classification and regression than traditional methods.Therefore,the support vector machine model recognition method is used to study the beam bridge damage.The main research work of this paper is as follows:(1)This article summarizes the background and significance of bridge damage identification and historical development by referring to a large number of domestic and foreign literatures.The identification method of bridge structure damage is introduced,especially the damage identification method under dynamic characteristics is introduced in detail,and the bridge damage index is firstly understood.The guidance is provided for pattern selection to select appropriate damage indicators.(2)This article introduces the pattern recognition method and theory and the types and principles of support vector machine.Provide the basis for selecting the appropriate support vector machine model for bridge damage level distinguish and bridge damage position distinguish in pattern recognition.Finally,the support vector machine classification machine is used to do Bridge damage location identification,and the support vector machine regression machine is used to do bridge damage degree identification.(3)The structural damage identification method based on modal vibration mode and curvature is deeply studied.Then,the simple support beam bridge is taken as an example to extract the damage index by finite element simulation method,and the sensitivity of the two indicators to bridge damage identification is carried out.In contrast,the curvature mode is obtained as the best indicator of damage identification as the input parameter of the support vector machine model.(4)Taking the continuous beam bridge as an example,the support vector machine classifier is used to analyze the damage location identification of the bridge.It is feasible to verify the support vector machine model based on the first four-order curvature mode as the input feature parameter.The comparison of the prediction accuracy of four different core functions yields the best core value.The RBF kernel function related parameters are optimized,and the results are analyzed and verified.