Study On Moving Force Identification Technology Of Hull Structure

There are potential safety hazards in the operation of ships and Marine engineering structures,so the health monitoring of structures is a hot topic in current researches.As a key link in structural health monitoring,moving force identification provides an important basis for structural safety assessment.However,in most cases,it is difficult to predict and measure the moving force directly.In the case that the direct measurement is not feasible,the identification of the moving force becomes an effective and low-cost method.Therefore,aiming at the moving force identification problem of hull structure,this paper mainly focuses on the moving force identification algorithm,and the main research contents are as follows:(1)Research on moving force identification technology based on time domain method was carried out.This paper expounds the basic principle of identifying the moving force with the time-domain method,takes the simply supported beam as an example to identify the moving force,and compares the identification accuracy of different measuring point positions,number of measuring points and different noise levels.The possible ill-posed problem of transfer matrix in time domain method is analyzed,and the improvement direction is proposed.(2)In order to solve the matrix ill-posed problem in time domain method,an improved method is adopted.Regularization parameters are selected by local convex curve criterion,and the moving force identification equation is solved by combining Tikhonov regularization method(Tikhonov＿Lcc).The feasibility of the algorithm is verified by taking the simply supported beam as an example,and the identification accuracy is analyzed under different noise levels,different number of measuring points and different locations of measuring points,and compared with the time domain method,more stable and accurate force values are obtained.(3)Traditional analytical algorithms are difficult to establish the relationship between moving force and response on complex structures.To solve this problem,a method for identifying moving force on ship structures based on BP neural network is developed.Firstly expounds the basic principle of BP neural network,studied the method of identification of moving force flow,then respectively with simply supported beam and the structure of hull grillage beam for example verifies the algorithm,in the final analysis the influence of different noise levels on the recognition accuracy,and noise into the training sample of the impact of network generalization,this paper discusses the different measuring point number affect the identification accuracy.The results show that the BP neural network algorithm can identify the moving force effectively.(4)Aiming at the efficiency problem of moving force identification,the extreme learning machine(ELM)algorithm is used to identify moving force on hull structures.Firstly expounds the basic principle of ELM algorithm,and explore how to carry on the moving force identification by the algorithm,and then with simply supported beam and the structure of hull grillage beam as example,comparing the precision of the identification results under different noise levels,in order to investigate the influence of noise into the training sample of algorithm,studies the influence of different number of measuring points on the identification results.Finally,compared with the BP neural network identification results,the results show that ELM algorithm has similar identification accuracy and higher identification efficiency,and analyzed the specific reasons for the improvement of identification efficiency.