• CN:11-2187/TH
  • ISSN:0577-6686

机械工程学报 ›› 2020, Vol. 56 ›› Issue (18): 116-122.doi: 10.3901/JME.2020.18.116

• 运载工程 • 上一篇    下一篇

扫码分享

形态学滤波方法在车轮非圆化信号降噪中的应用

陶功权, 谢清林, 刘晓龙, 任德祥, 温泽峰   

  1. 西南交通大学牵引动力国家重点实验室 成都 610031
  • 收稿日期:2019-09-30 修回日期:2020-03-30 出版日期:2020-09-20 发布日期:2020-11-17
  • 通讯作者: 温泽峰(通信作者),男,1976年出生,博士,研究员,博士研究生导师。主要研究方向为轮轨关系和振动噪声。E-mail:zfwen@home.swjtu.edu.cn
  • 作者简介:陶功权,男,1989年出生,博士,助理研究员。主要研究方向为轮轨关系。E-mail:taogongquan@swjtu.edu.cn
  • 基金资助:
    国家重点研发计划(2016YFB1200501-005)、国家自然科学基金(51875484)、四川省科技计划(2019YFH0053)、牵引动力国家重点实验室自主课题(2020TPL-T03、2020TPL-T12)和中央高校基本科研业务费专项资金(2682020CX52)资助项目。

Application of Morphology Filtering Method in the De-nosing of Wheel Out-of-roundness Signals

TAO Gongquan, XIE Qinglin, LIU Xiaolong, REN Dexiang, WEN Zefeng   

  1. State Key Laboratory of Traction Power, Southwest Jiaotong University, Chengdu 610031
  • Received:2019-09-30 Revised:2020-03-30 Online:2020-09-20 Published:2020-11-17

摘要: 铁道车辆车轮踏面通常存在麻坑等缺陷,在车轮非圆化测试信号中往往包含这些干扰信号,其尺寸小于轮轨接触斑尺寸。采用形态学滤波方法对车轮非圆化信号进行降噪处理。研究不同尺度和不同形状结构元素对车轮非圆化信号的降噪效果。针对形态学滤波在信号"细节刻画"上较为粗糙的缺点,构建形态学滤波+三次样条平滑的降噪方法,并且对比新方法与传统的车轮非圆化信号降噪方法的降噪效果。以不同降噪方法得到的车轮非圆化信号作为输入,研究车轮非圆化对轮轨动力相互作用的影响。结果表明,采用幅值为1 mm宽度为4 mm的三角形结构元素进行形态学滤波+三次样条平滑的降噪方法能更好地滤除车轮非圆化信号中的噪声信号,滤波效果优于传统的车轮非圆化信号处理方法,并且能够保留原始信号特征。

关键词: 车轮非圆化, 数学形态学, 滤波, 降噪, 轮轨力

Abstract: Wheel defects, such as pitting, generally occur on the tread of railway vehicle wheels. The measured wheel out-of-roundness (OOR) signals usually contain interference signals induced by pitting or other defects. The dimension of pitting is less than wheel-rail contact patch. The morphology filtering method is used to de-noise the wheel OOR signals. The influence of structure element's dimension and shape of morphology filter on the de-nosing of wheel OOR signals is investigated. A new de-nosing method combining morphology filtering with cubic spline smoothing is proposed to overcome the shortcoming of morphology filter. The results of wheel OOR signals processed using the new and ordinary methods are compared. Moreover, the influence of wheel OOR on wheel-rail dynamic interaction is also investigated by using wheel OOR signals from different de-nosing methods. The results indicate that the morphology filter with triangular structure element, of which the amplitude and width are 1 mm and 4 mm, respectively, combined with cubic spline smoothing is more suitable in the de-nosing of wheel OOR signals. The results of this method applied to process wheel OOR signals show that the signal-noise ratio of the processed signal is much higher than the ordinary method and can keep the main characteristics of the original signal.

Key words: wheel out-of-roundness, mathematical morphology, filter, de-noising, wheel-rail force

中图分类号: