Research On Structural Damage Identification Method Based On Improved Optimization Algorithm And Limited Measurement Data

Affected by the external environment and abnormal load,engineering structures and mechanical equipment inevitably suffer from structural damage such as bolt loosening,surface abrasion and crack,which leads to increasingly prominent safety problems.It is particularly important to timely monitor the structural health status information.As an important part of structural health monitoring,structural damage identification aims to accurately locate and quantify damage.In this thesis,beam,plate and rail structures are taken as the research objects,and structural damage identification is studied based on wavelet transform and optimization algorithm.The main contents and conclusions are as follows(1)The modal curvature is calculated by the structural mode shape and approximate calculation formula of central difference,and the suspicious damage elements are located by wavelet transform.Then,based on the optimization algorithm and the damage quantification objective function in the form of the product of RMSD value of structural modal shape and natural frequency,the suspicious damage elements are further optimized to identify the exact location and extent of true damage elements.(2)An improved algorithm(CMVPA＿HJ),which combines MVPA optimization algorithm with TLBO,chaos optimization and Hooke-Jeeves local search algorithm,is proposed.Compared with four other optimization algorithms,such as MVPA,GAPSO,PSO and GA,CMVPA＿HJ is applied to the damage identification of beam,plate and rail structures.(3)From simply supported beam,cantilever plate to three span simply supported rail model,the proposed damage identification method is verified by numerical simulation and experimental verification.The influence of different noise level,different optimization algorithm,different objective function of damage quantification on the effect of damage identification is analyzed.The optimal sensor placement problem is transformed into an optimization calculation problem.The objective function of sensor placement optimization is constructed based on the sensitivity of vibration mode to structural damage,and then CMVPA＿HJ is used to solve the optimal location of a given number of sensors,and the damage identification effect of the measurement point is verified.The results show that the method based on wavelet location and optimization algorithm can realize the damage identification of beam,plate and rail structure.Test function verification results show that CMVPA＿HJ is better than the other four optimization algorithms,which has faster convergence speed and better convergence value.The damage location method based on wavelet transform and modal curvature can effectively locate the damage location,even in high noise level and limited number of measuring points.In the combination of different optimization algorithms and objective function of damage quantification,the results show that the damage quantification objective function of vibration mode in the form of RMSD and CMVPA＿HJ has the best effect on the damage quantification,not only the damage quantification speed is fast,the number of damage misjudgment element is small,but also the calculation error of damage extent is smaller.According to the optimization algorithm and the objective function of sensor placement optimization,the location of sensor placement can be calculated.Even if the number of measuring points does not exceed 50% of the total number of measuring points,the proposed damage identification method can still effectively identify the structural damage.