基于轮轨噪声的钢轨裂纹故障诊断方法研究

项目名称： 基于轮轨噪声的钢轨裂纹故障诊断方法研究

项目编号： No.11502198

项目类型： 青年科学基金项目

立项/批准年度： 2016

项目学科： 数理科学和化学

项目作者： 高瑞鹏

作者单位： 西安理工大学

项目金额： 22万元

中文摘要： 高速重载列车钢轨产生裂纹的概率大，其引发的安全隐患越来越突出，导致维修成本逐年显著递增，急需一种先进有效、易于大规模使应用钢轨裂纹检测技术。本项目拟在前期研究基础之上，针对现有检测技术受速度限制、不利于大规模推广应用不足，结合现代信号处理技术如压缩感知技术，分析轮轨噪声-钢轨裂纹力学关系，建立其数学模型，进行算法优化及代价函数设计，提出一种准确度高、性能优越的钢轨裂纹无损检测算法；通过对现场环境信号的分析并辅以智能算法建立声源叠加模型，以此为基础使用独立分量进行信号分离，设计效果更好的信号预处理算法；同时以Twins软件包为依托搭建能够高速在线处理的系统硬件平台。本研究对于立高效的钢轨裂纹无损检测，具有重要的科学意义和实用价值。

中文关键词： 轮轨噪声；故障诊断；信号分析；钢轨裂纹

英文摘要： There is a high probability of rail crack by high speed and heavy load train, the safety hidden trouble caused more and more prominent, which is resulting in that the repair costs significantly increased year by year. Therefore a rail crack detection technology which is dvanced and efficient and easily to promote to use of large-scale is needed. The project which is on the basis of previous research, is in view of the existing detection technology of limit which is not conducive to the lack of large-scale promotion and application. It establishes mathematical model of wheel / rail noise - rail crack which is combined with modern signal processing technology such as compressed sensing technology and analysis the relations of wheel / rail noise - rail crack mechanics. It will propose an algorithm which is high accuracy, excellent performance of rail nondestructive crack detection through optimization and cost function design. It will design a better signal preprocessing algorithm which is using signal separation of independent component through the analysis of the field environment signals and supplemented by intelligent algorithm, the establishment of sound source superposition model. At the same time, it will also use the Twins software package to build the hardware platform system capable of high speed online processing. The study of nondestructive testing for the rail crack made efficient, it has important scientific significance and practical value.

英文关键词： The wheel / rail noise;Fault diagnosis;Signal analysis;Rail crack