Study On Crack Damage Identification Of Track Plate Based On Vibration Response

The aim of this study is to identify the degree of track plate cracks of ballast track structure by using the method of feature extraction and pattern recognition to provide the basis for making effective and reasonable maintenance and repair decision.According to the common crack damage form and structural characteristics of ballastless track,combined with the development of damage identification technology,the damage characteristics index is extracted by using a variety of data processing methods based on structural vibration response and combined with the modern pattern recognition technology to realize the crack degree of ballastless track.The main works accomplished are as follows:(1)According to the characteristics of the track structure and damage type,a variety of damage identification technology is compared and the superiority,limitations and applicability are analyzed,the damage identification applied for ballastless track structural cracks is selected.(2)According to the common crack form of track structure,the structural force characteristic is analyzed and various damage simulation methods are demonstrated,finally,the simulation method of elastic modulus is established,and the finite element simulation model with different degree of damage is established.The detection points in the different parts of the structure are set and the vibration response of each detection point is extracted by applying impact load on the finite element model.(3)The vibration response signals of different parts of the structure are analyzed by using the means of signal analysis in the time domain,frequency domain and time-frequency range to excavate the regularity information of the degree of damage and extract the characteristic index.(4)In order to fully reflect the structural characteristics,the data correlation theory and the vibration transmission characteristics are used to analyze the variation degree of the correlation degree and the vibration transmission characteristic of different parts with the damage degree in the track structure,and further extract the characteristic index.(5)From the perspective of reducing the data dimension,the principal component analysis is carried out and form the perspective of increasing the sample capacity,the pattern recognition method of neural network is used to identify the degree of crack damage on the basis of the established characteristic index.