Modal Parameter Identification And Damage Diagnosis Of Offshore Platform Structure Based On Monitoring Data

Offshore platform has become an indispensable and important infrastructure for offshore oil and gas exploration and exploitation.Therefore,it is of great significance to study the method of offshore platform structure health monitoring for timely detection of damage and protection of life and property safety.The purpose of this paper is to realize the modal parameter identification and on-line damage diagnosis of the offshore platform based on the monitoring data through the indoor vibration monitoring of the offshore platform structure.Based on the indoor platform modal,the offshore platform vibration monitoring system is developed to obtain the vibration monitoring data of undamaged and damaged structures.Then the modal parameters of the platform model are obtained by using the enhanced frequency domain decomposition method and the stochastic subspace identification method respectively.The modal identification results of the two methods are compared.Finally,the damage index is established based on the flexibility matrix to locate the structural damage.The results show that the enhanced frequency domain decomposition method can identify more modal information,while the stochastic subspace identification method is more sensitive to damage changes.In order to solve the problem of incomplete modal information recognition in the stochastic subspace identification method,the variational mode decomposition method was introduced to decompose and reconstruct the signals,and the particle swarm optimization algorithm was used to search the optimal combination values of the variational mode decomposition method.The simulation signals were established and compared with the results obtained by empirical mode decomposition and ensemble empirical mode decomposition.Finally,variational modal decomposition method and stochastic subspace identification method are combined to identify the modal parameters of the platform.The results show that the modal recognition method based on variational modal decomposition and stochastic subspace identification method can identify more order modes,and the modal recognition results are stable compared with the modal parameters calculated by single stochastic subspace identification method.In order to avoid damage misjudgment caused by modal information loss caused by algorithm difference,Hankel matrix and chi-square test methods in stochastic subspace identification method were used to establish damage index based on Hankel matrix residuals before and after structural damage.On the basis of chi-square test,Mahalanobis distance method was introduced to update the damage index.The results show that the damage index based on Mahalanobis distance has good robustness to environmental excitation and has the capability of structural health monitoring.Finally,the modal parameters of the structure are calculated according to the Hankel matrix of the system,and a method of calculating the relative change value of the mode shape is proposed to complete the damage localization of the platform.In order to realize the online monitoring and damage diagnosis function of the platform,based on the Alex Net model in the convolutional neural network,the stochastic subspace identification steady state graph is used as the input sample,and the Benchmark model is simulated under three working conditions,and good results are obtained.Then it was applied to indoor platform damage diagnosis,and the variational mode decomposition method was used to decompose and reconstruct the signal adaptively and update the steady-state diagram.The results show that the improved neural network test accuracy is improved,and the test results are more stable,which can be applied to the health monitoring of offshore platform.