轮对复杂激扰下高铁轴箱轴承故障振动特征提取与演变规律研究

Axle-box bearings of high-speed railway are always running at a high speed in the environment with high-load and strong impact. They are disturbed by strong noise and affected by significant coupling vibration, especially the uncertain complex excitation of wheel-set. All of these lead to changes in vibration characteristics and characteristic evolution law of axle-box bearings, which have become the key points and difficulties for fault defection and state prediction of axle-box bearing in real world. Firstly, the disturbance effect of typical wheel faults and common wheel-set deviation states will be studied by setting the independent models of wheel-set and axle-box bearing based on the vehicle-rail dynamic analysis model. Secondly, blind separation method for nonlinear and non-stationary multi-source mixed signals will be explored to peeling off complex excitation of wheel-set and restoring the real vibration source information of axle-box bearing. The stochastic resonance method will be used to enhance the weak feature, and then the fault features of axle-box bearing can be extracted. Finally, the dimensionless stability representation method based on sparse characteristics for vibration signals of axle-box bearing with faults will be studied, and fault characteristics parameter base will be established by combining the vibration information. The evolution law of fault characteristics of axle-box bearing under complex excitation of wheel-set will be revealed from multiple perspectives. It is of great significance to health monitoring, condition prediction, safety warning and condition-based maintenance for axle-box bearings in high-speed railway that to grasp the mechanism of wheel-set multi-state influence on axle-box bearing and to solve the problems of fault feature extraction and evolution law establishment for axle-box bearing under complex service condition with dynamic excitation.

高铁轴箱轴承处于高转速、大载荷、强冲击的运行环境，受强噪声干扰且耦合振动效应显著，特别是轮对的不确定复杂激扰作用，导致轴箱轴承振动特性及特征演变规律发生改变，成为高速列车运行条件下轴箱轴承故障检测与状态预测的关键点和难点。通过在车辆-轨道耦合动力学分析模型基础上对轴箱轴承及轮对分开建模，研究车轮典型故障、轮对常见偏差状态对轴箱轴承的激扰作用；突破非线性、非平稳多源混叠信号的盲分离方法，剥离轮对复杂激扰还原轴箱轴承振源信息，利用随机共振方法，增强特征，实现轴箱轴承故障特征提取与诊断；研究基于稀疏特性的无量纲的故障状态信号稳定性表征方法，结合振动信息量建立故障特征参量库，多角度揭示轴箱轴承故障特征在轮对复杂激扰下的演变规律。掌握轮对多状态对轴箱轴承的影响机制，解决复杂激扰下轴箱轴承故障特征提取及其演变规律表征与建立问题，对高铁轴箱轴承故障检测、状态预测、安全预警及状态修的实现具有重要意义。

本项目针对高铁轮对轴承故障特征提取与状态监测难题，从复杂激扰下轮对轴承振动特性、车轮对轴承的影响机制及轮对轴承故障特征提取等三个方面，对轮对轴承故障诊断及复杂环境下故障特征演变规律两个层次的理论、方法、关键技术开展研究。经过三年的努力，取得了如下成果：（1）基于轴箱与构架、构架与车体的动态耦合方程，建立了包含轴箱系统的车辆—轨道耦合动力学模型。（2）提出了一种有效的轮对轴承振动激励模型，获取了轮对轴承幅频曲线，并建立幅频曲线与车轮多种典型故障的映射关系。（3）揭示了车轮多边形、车轮扁疤等复杂激扰对轴箱系统的影响规律，高阶车轮多边形磨损引起的轴箱垂直振动加速度的频域主要为400～600 Hz，高阶车轮多边形磨损（17～21阶）对轴箱系统的影响比低阶车轮多边形磨损（1～5阶）更为显著；当列车运行速度在250 km/h左右时，车轮扁疤长度需限制在30 mm以内。（4）在共振频率探测方面，从多级频谱分割、信号分离算子、自适应变分模态分解等方法入手，建立了轮对轴承复杂振动信号降噪与频带自适应划分技术框架，实现了轮对轴承共振频带尽完备分离与探测，相较于快速峭度图方面等方法，本项目研究成果对紧密性频带分离具有显著的应用效果。（5）为了进一步提取轮对轴承故障信息并表征其在复杂环境中的演变规律，本项目研究了一个新的改进的相关广义Lp/Lq范数指标，提高了鲁棒性和对脉冲的探测能力；在平方包络谱中定义了故障特征的信噪比，并使用包含故障特征频率及其倍数的循环频带来替换循环频率，研究了一种考虑了速度波动影响的平方包络谱谐波噪声比作为目标函数的新盲解卷积算法。本项目利用高速列车轮对跑合试验台轮对轴承故障实验及实际运行条件下的轴箱振动信号对上述研究成果进行了验证与优化，共发表了论文20篇，包括SCI期刊论文18篇、EI期刊论文2篇，完成博士学位论文2篇，完成硕士学位论文4篇，申请国家发明专利6项。