Research On Bridge Damage Identification Method Based On Support Vector Machine

Bridge structure will be affected by various loads and environmental factors during operation.Over time,the performance may degrade,which makes the structure face safety risks.In order to ensure the safety of bridge structure operation,it is necessary to evaluate the damage state of the structure and accurately grasp the operation state of the bridge.It is a hot issue in the field of bridge health monitoring to identify the damage,find the damage and determine the location and degree of the damage.Based on the support vector machine method,this paper identifies the damage location and damage degree after simulating the damage of bridge structure in the operation process.The research work includes the following aspects:(1)This paper systematically introduces the related theories of support vector machine,and deduces the algorithms of support vector classification machine and support vector regression machine,which provides the basis for the application of support vector machine in structural damage identification.(2)This paper summarizes the damage identification methods based on natural frequency and vibration mode,compares and analyzes the basic theories of three damage identification indexes based on frequency change and four damage identification indexes based on vibration mode change,and evaluates the advantages and disadvantages and application scope of several indexes,which provides a theoretical basis for the application of support vector machine method with dynamic response as input in damage identification.(3)The damage identification method based on support vector machine and the effectiveness and applicability of various damage identification indexes are studied.Taking the concrete simply supported beam bridge as the object,according to the demand of each damage identification index,the corresponding dynamic response eigenvector is constructed,and then the damage location and damage degree are identified by using the support vector classification machine and support vector regression machine,and the good recognition effect is obtained,and the anti noise performance of each index is further studied.Comparing the damage identification effect and anti noise performance of different indicators,it is found that for damage location identification,the three indicators based on frequency change have better identification effect and stronger anti noise performance;for damage degree identification,the identification effect of frequency relative change is better than the indicators,and the anti noise performance is the strongest.(4)On this basis,in view of the shortcomings of the existing indicators,two new damage identification indicators are proposed,and the identification effect and anti noise performance of the new indicators are studied.It is found that in the two new indicators,the effect of the index identification obtained by the normalization of the frequency relative change ratio and the vibration mode ratio is good,the promotion ability is strong,and the anti noise performance in the damage location identification is better Combining the two indicators in an appropriate way can realize the complementary advantages of the two indicators.Based on the support vector machine method,this paper uses structural modal parameters to identify the location and extent of damage to the bridge structure,and obtains a relatively ideal effect,which verifies the feasibility and effectiveness of the damage identification method based on the support vector machine.It provides technical support for structural damage identification based on the support vector machine method,and provides a technical reserve for ensuring the safe operation of bridges.