Traceable Correction Method for the Online Measurement System of Train Wheelset Profile Based on Neural Network

Abstract

Online measurement system of train wheelset profile can automatically measure the wheelset size of all trains passing through the detection zone and timely pre-alarm the oversized wheelset to ensure the safe and smooth operation of the train. In this paper, the main technologies of wheelset size online detection at home and abroad are retrieved, and the problems are analyzed, such as data accuracy is difficult to maintain for a long time, the detection index is imperfect, and the dynamic and static calibration of the system has different in practical applications. Also through analyzing the main sources of detection error from the detection principle and application conditions, expounding the influence of radial runout, and axial displacement on measurement results emphatically, a parameter tracing method based on neural network is proposed and key calibration steps and engineering application examples are also given.