微弱信号的混沌同步检测机理及地铁车辆轮对故障的早期诊断

Due to the metro vehicle complex work conditions and a large amount of noise, traditional signal detection methods are difficult to achieve early fault diagnosis of metro wheels. This is a security risk to the metro operation. To this end, the project studied a new method of weak signal detection based on chaotic synchronization theory. The method, which has important scientific significance and application value, was used to early fault diagnosis of metro wheels. It was different from the conventional detection principle. The project analyzed characteristics of weak fault signals when the metro vehicle was running, constructed a weak signal chaotic synchronization detection system to detect wheels’ faults. Then, put the parameter identification method and adaptive algorithm, analyzed synchronization detection characteristics of vehicle wheels’ early weak fault signals, tested detection capability and feasibility of the new method to common faults of traction gears and axle box bearings in vehicle wheels. The project has the following innovations: first, the new method has high detection speed and avoid solve complex phase space trajectories, which is needed in the existing chaotic detection methods. Second, has strong anti-interference ability, even in strong noise environment it can be completed accurately detect weak signals. Third, the detection device is easy to implement, which can be build by the existing basic chaotic systems. The project will further promote the study of the mechanism of weak signal detection based on chaotic synchronization. And provide a new theory and a new method for early diagnosis of metro vehicle wheels failure.

地铁车辆运行时工况复杂且噪声干扰严重，传统的信号检测方法难以实现车辆轮对故障的早期诊断，给运营带来安全隐患。为此，本项目基于混沌同步理论，研究微弱信号检测的新方法，应用于轮对故障的早期诊断，具有重要的科学意义和应用价值。与常规检测原理不同，本项目分析地铁车辆运行时轮对故障微弱信号的特征，构建轮对故障微弱信号混沌同步检测系统，提出相应的参数辨识方法和自适应算法，并进行车辆轮对故障早期高精度混沌同步检测特性分析,验证其对车辆轮对中牵引齿轮和轴箱轴承常见故障的检测能力和可行性。项目有如下几个创新之处：一是检测方法实时性强，避免了现有混沌检测方法对相空间轨迹的复杂求解；二是抗干扰能力强，即使在强噪声环境下也能完成准确检测；三是检测装置易于实现，利用现有基本混沌系统即可搭建。本项目研究成果将推动混沌同步检测微弱信号的机理研究，为地铁车辆轮对故障的早期诊断提供一种新理论和新方法。

本项目以强噪声背景下的微弱信号为研究对象，研究了基于混沌同步理论进对其进行预处理和检测的新方法，并将该方法应用于车辆轮对振动信号的前期预处理和故障判定。项目的研究内容基本按原定计划进行，主要完成了以下工作：. （1）分析了基于混沌系统的微弱信号检测机理。提出了利用Duffing振子输出值的方差峰值结合遗传算法检测淹没在强噪声背景中频率未知微弱信号的新方法。该研究解决了现有混沌振子类检测方法必须已知信号频率的限制，能准确、快速检测待测信号频率。. （2）研究了基于混沌同步系统的微弱信号检测机理。设计了混沌同步方式检测微弱信号的系统框架结构，研究了系统的稳定性，分析了系统参数对同步状态和运行轨迹的影响。并提出了采用MUSIC算法处理同步误差得到频率值，再将产生的同步参考信号与待测信号进行相关运算得到相位差，最终通过自适应控制器得到幅值估计值的方法。此研究利用了混沌系统可同步的特性，保留了噪声免疫性和初值敏感性。在构建成本、参数设定、检测效率等方面有很大改进，提高了适用性，为其应用于振动数据的实际处理创造了条件。. （3）搭建了基于混沌同步系统的实际检测电路和原理性样机，并与现有混沌类检测方法进行对比。以Chua电路、Lorenz系统和Duffing系统为典型，进行了模型搭建、仿真分析和实验验证，并与功率谱等传统频域分析技相结合，进一步简化了信号处理的复杂程度。. （4）基于所构建混沌同步信号检测系统进行了车辆轮对轴承故障早期诊断的研究。从轴承的结构出发，分析了不同部位发生故障后所对应的故障特征频率，将发生故障时的实际振动数据作为待测信号接入搭建好的混沌同步系统进行处理，并与传统检测方法得到的检测数据进行了对比分析。. 上述研究均通过理论分析、数值仿真模拟和搭建原理性样机的方式进行了检验。而强噪声背景下微弱信号的检测方法在实际工程中有着大量的需求，所研究的内容具有广阔的应用前景和工程应用价值。