Research On Damage Identification Of Ship Structure Based On Optical Fiber Sensor

In recent years,the development of modern technology,especially in the structural health monitoring technology updates,ships,spacecraft and other complex structures to extend the service life of the high demand.Fiber Bragg grating(FBG)sensor has the advantages of easy installation,high space utilization,and high durability under extreme conditions,and is increasingly used in damage detection.In particular,damage detection in the fields of ship structures and aerospace vehicles has received widespread attention from researchers.According to the needs of the inspection object and the inspection environment,the design of a multi-source impact position identification and multi-point crack detection system based on FBG sensors has an important research value for the health monitoring of complex ship structures.Aiming at the problems of complex design,low timeliness,unfavorable to universal application and unsuitable for multi-point inspection of the impact location and damage detection system of ship structure,a hyperbolic localization method using FBG sensor combined with fast blind source signal separation is proposed.Firstly,the mixed signals generated by multi-source impacts are pre-meaned and whitened,and the collected mixed signals are separated with the criterion of negative entropy maximization theory to determine the number of impact sources,calculate the arrival time delay of each independent signal,and then use the hyperbolic localization model to determine the location of impact sources.The existing acoustic emission signal-based crack identification target is single,and the target accuracy is not high in the presence of multiple cracks,and the identification features are few.A method using strongly weighted center of mass combined with particle swarm is proposed to deal with the problem of pseudo-crack sources,and the signals other than cracks are removed.The strain signals detected by FBG sensors at the time of impact or crack damage generation are used to determine the characteristic information of crack sources such as length and angle by an improved particle swarm algorithm based on the rejection of pseudo-crack sources.The range of optimal learning factors was simulated by means of reliable simulations before the experiments.The experimental platform built by simulation was used to conduct several experiments on multi-point impact and crack identification of ship structures.The data of the multiple experiments were averaged,and the average localization accuracy of the obtained multi-source impact localization reached 1.95 cm with a localization delay of about25 ms.In addition,the simulation results of the improved particle swarm optimization algorithm show that the average number of iterations is 16,the average error is 1.85,the average time spent for crack parameter identification and localization is 1.75 s,and the value of learning factor is not less than 0.4.It provides a feasible method for the detection of cracks in ship structures.