Research On Wheel Set Detection Technology For Rail Vehicles Based On Dual-line Structured Light

With the continuous development of China’s rail vehicle business,high-speed rail,moving train,subway,light rail and other railroad transportation industry to reach the unprecedented prosperous scene,the wheelset is an important support and walking parts of the rail train,so its geometric parameters and tread quality inspection of the safe operation of the vehicle has great significance.According to the requirements of rail vehicle wheelset manufacturing quality inspection,the traditional manual measurement error is large,and the wheelset measurement and evaluation methods based on points and lines cannot fully reflect the whole design,manufacturing and usage information of the wheel pair.The wheelset automatic inspection system mainly includes drive mechanism,positioning support mechanism and inspection control system.In this system,the axial and radial positioning of the wheel pair under test is completed through the positioning support mechanism in the form of imitation V-block and alignment control with multi-sensor information fusion,and the positioning error of the system is given error compensation;the friction torque and clamping force required to rotate the wheelset are deduced from the drive mechanism,the clamping force is provided by electric cylinder,and the wheelset is driven by servo motor to rotate;the high-precision ball screw slide is selected to carry the line structure optical sensor and laser distance measuring sensor to collect the profile information of the wheelset,and the synchronous control system is designed to obtain the point cloud data of the wheelset.The dual structured light measurement method and principle are introduced to build a dual line structured light 3D measurement system,which can solve the problem that a single line structured light sensor cannot collect complete tread information.Calibration of the measurement system and analysis of measurement errors.The collected wheel pair point cloud data is de-noised by a straight-through-statistical filtering algorithm,and the two point clouds are aligned by FPFH coarse alignment and ICP fine alignment to achieve data fusion,then the point cloud data of the inner surface of the wheel is extracted to fit the inner reference surface,and the overall wheel point cloud is rotated according to its angle with the coordinate axis plane.Based on the curvature features,the geometric key feature points are extracted,weight factors are determined,data points are parameterised and node vectors are solved to invert the control vertices of the entire tread surface,and the NURBS surface interpolation of the wheel pair tread surface is reconstructed to restore the measured wheel pair tread surface profile as far as possible.The reconstructed tread surface was analysed in Geomagic Qualify and Solid Works for reconstruction model errors and smoothness.Based on the inherent geometric relationships defined by each parameter of the wheel tread,the inner reference surface of the wheel is first fitted,and the target geometric parameters of the wheelset are extracted in the 3D reconstructed model according to the definition of each parameter.The experimental results prove that the measurement accuracy of wheel diameter and rim width is ±0.2mm,and the measurement accuracy of rim width and height is ±0.25 mm,and the error of QR value is less than 0.25 mm.The measurement results meet the measurement accuracy requirements.It provides a new solution for the static inspection technology of wheel dimensional parameters of rail trains.