Multi-sensor Detection Method For Crack Acoustic Emission Signals

With the rapid development of economy of our country in recent years,the speed of development of high-speed railway is increasing day by day,at the same time,with the rapid development of high-speed railway,the safety of railway transportation attracts more and more attention by society and scholars in the related field.This paper starts with the safety of railway transportation,researches the damage detection of railway,and makes a series of testing and simulation on some methods of railway nondestructive testing,and improves some of these methods.This paper researches multi-sensor information fusion method used in the field of railway nondestructive testing.Firstly we choose adaptive fusion algorithm and Kalman filtering algorithm to deal with the data,then we choose blind source separation algorithm according to the deficiency of the algorithms above.The signals in this paper are collected from experiment table built in laboratory,the experiment table uses steel board to instead rail,rub on the steel board to produce noise,at the same time crack a pencil and collect the crack signal to get simulated signal of rail fracture.In this paper the process of signal processing has two ways,one is to fuse the signal at first and then pass Kalman filter to get breakpoint signal,the other way is using blind source separation algorithm to deal with the signal of multi-sensor,and get the breakpoint signal.During the experiment we find that traditional blind source separation algorithm can't control the result we want,so we propose the improvement method in preprocessing and improve the blind source separation algorithm into blind source extraction algorithm,and then compare the result before and after improvement to explain the effect of improvement.At last the paper discusses the method to solve the question of undetermined problem because of few number of sensors in real signal collection,and decides to use Ensemble Empirical Mode Decomposition algorithm to solve this problem.Then this paper utilizes this algorithm to resolve the simulated rail crack signal in the actual operation of the train and makes the signal adaptive.At last this paper analysies the signal,the result proves the feasibility of improved blind source extraction algorithm.