Research On Automatic Detection For Bridge Cables Based Multi-Sensor Data Fusion Using The Snake-Like Robot

Cable-stayed bridge is the most important bridge of large span bridge type, the cablesystem of cable-stayed bridge is the main load-bearing structures and power transmissioncomponents, but the rope exposes long-term and is eroded by sunshine, rain etc. If it hasdamaged, it will reduce the use and durability of the structure. And once broken cableaccident occurs, it will cause a very heavy economic losses and casualties. So healthmonitoring and maintenance is particularly important on the ropes.In order to achieve the automatic non-destructive testing of the cable-stayed bridgecables, this paper researches the automatic non-destructive testing methods of cable-stayedbridge, and puts forward the automatic detecting method which based on snake-like robot andmulti-sensor data fusion. Using the snake-like robot spiral climbing motion which carriesmulti leakage magnetic flux sensor, it realizes on-line automation detection. The use of datafusion technology to automatically detect the multi-sensor data fusion bridge cable defects,respectively, from the data level, feature level, decision-making level to achieve all fusionalgorithms. The main article content including:(1) Design the automatic detection system based on the snake-like robot for bridgecables defects.According to the bionics background of the biological snakes to introduce the snake-likerobot body structure design and control system design, focus on the snake-like robot spiralclimbing form which applies in the defect inspection of the bridge cables, and design of thesnake robot to achieve a variety of typical motion forms.(2) Proposed magnetic flux leakage nondestructive testing technology in the defectinspection of the bridge cables.Firstly, introduce bridge cables structure,testing content and type of defect; Secondly，followed by a detailed description of magnetic flux leakage nondestructive testing technology;Finally, use the Ansys software simulation to research the relationship between of bridgecable defects parameters and magnetic flux leakage signal.(3) Analysis bridge cable-based multi-sensor data fusion defects automatic detectionmethod. Based on theoretical knowledge, in the data layer using a weighted average signal fusion,in the characteristic layer using support vector machine as defect classification andidentification platform, in the decision layer using D-S evidence theory to make final decision,using Matlab to verify the data fusion methods. The experimental results show that the use ofdata fusion algorithm can not only reduce the uncertainty of the system, but also caneffectively improve the recognition accuracy and reliability of rope defects.