Research On Common Fault Detection System Of Pantograph Head Based On Structured Light

The pantograph is an important part of the power supply system of the electrified railway.It is an important power take-off device in the operation of the electric locomotive.The pantograph fault mostly occurs in the pantograph head.The common failure modes include the contact strip wear out of limit,the deviation of the centerline,the abnormal posture of the pantograph head and the deformation of the horns,etc.If the pantograph fault is not handled in time,serious traffic accidents may be caused.It is an inevitable choice to detect the fault of the pantograph head to ensure the safety of railway transportation.The traditional methods of pantograph contact strip wear detection have some problems,such as low accuracy,heavy workload and need to stop for detection,etc.;the research on other fault detection methods of pantograph head is relatively less,which mainly focused on image method,and the detection stability is insufficient.The development of 3D measurement technology provides a new solution for industrial detection.After comparing different 3D measurement methods,combined with the actual requirements of pantograph head fault detection,the line structure light measurement method is selected to complete the detection,to improve the detection efficiency and stability.This thesis focuses on related research on the hardware design of the pantograph head fault detection system,pantograph head fault detection algorithm design and experimental verification,etc.The fault detection system of the pantograph head was designed.The hardware composition,automation workflow,initial collection setting,and anti-interference design of the system were introduced in detail to realize the automatic and high-precision detection of pantograph head fault.Combined with the system installation environment,passing speed and other factors,the key components of structured light sensors such as camera and laser are selected.At the same time,the measurement model of structured light,the calibration method and the algorithm of extracting the centerline of the light strip were described in detail,and the sensor calibration and accuracy verification were carried out to meet the needs of highspeed and long-distance measurement of the system and obtain enough scanning contour of pantograph head.Because a single structure light sensor can’t cover pantograph head of different sizes,a multi-sensor combined measurement scheme was designed to realize data collection of pantograph head point cloud from different perspectives;considering the characteristics of sensor installation and data types,the point cloud denoising and smoothing based on radius filter and Gaussian filter were carried out,also the plane rough splicing–ICP fine splicing algorithm was proposed to realize the point cloud splicing from different perspectives and complete the point cloud preprocessing.Subsequently,through the sawtooth target splicing experiment,the accuracy of point cloud splicing was verified,and the pantograph point cloud panoramic imaging is realized.To give scope to the advantages of 3D measurement,analyze the pantograph structure,select the contact strip bracket as the calculation basis,and design the contact strip wear detection algorithm based on the bracket.According to the two scenarios of the contact strip bracket with no curvature and curvature,the straight bracket fitting algorithm based on RANSAC–LM and the curved bracket fitting algorithm based on Whittaker smooth were proposed to reduce the fitting error of bracket and improve the accuracy and stability of the contact strip wear detection.Combined with the advantages of pantograph head panoramic point cloud imaging,the detection algorithm of centerline offset,pantograph head posture,and horns deformation based on point cloud data were proposed to achieve more comprehensive pantograph head fault detection.Finally,using the experimental platform to simulate the pantograph head fault and the historical data collected in the field,the accuracy of the pantograph head fault detection algorithm was described qualitatively and quantitatively,which proves the feasibility and advanced nature of the system.The results show that the system can meet the requirements of practical application,and has the advantages of good stability,high precision,and non-stop detection.At the same time,the system error analysis was carried out to indicate the direction for the follow-up system optimization and improvement.