基于发动机磨粒SECM-SICM图像的故障诊断方法的研究

Because debris carry work states of engine,debris analysis is an important method of engine fault diagnosis. Due to limitation of the detection device, only size, 2D shape and chemical composition of debris are usually used. Neither 3D micro-image of debris nor the electronic and chemic distribution of debris is used at present.However, 3D micro-image of debris can uncover the cause and processing of debris and the distribution of faradic current and chemical composition can reveal producing environment of debris. If 3D micro-image and electrochemical distribution information are hanged together, not only the static information of debris, but also the dynamical procedure of debris can be revealed. So the work state and the fault cause of engine can be acquired effectually. This study proposes to get the fast acquisition method of SECM - SICM image of debris by the high frequency vibration holder of scanning electron microscopy (SEM) and the trajectory planning. As a result, a lot of images about the 3D surface morphology and electrochemical distribution on the debris desquamating surface are obtained as the necessary data sets for engine fault diagnosis. The method of multi-objective optimal fault diagnosis is designed in this project. Complex mapping relationships are decoupled by selecting appropriate feature sets, researching multi-objective functions and designing the algorithm structure. Finally, engine fault diagnosis method based on SECM-SICM images of debris is presented to improve the correctness and accuracy of engine fault detection and diagnosis.

磨粒携带着发动机重要的状态信息，是其故障诊断的一个重要手段。由于受到检测技术的限制，目前只利用了磨粒大小、二维形状等，未涉及更微观层次的结构和电、化学分布相关信息。磨粒剥裂面的三维微观形貌反映了磨粒产生原因；剥裂面的感应电流分布和化学成份特性则记录了磨粒产生环境。两者的结合不仅能够展现磨粒产生的静态原因，还能够呈现其产生的动态过程，进而有效地揭示发动机的工作状态和故障成因。本项目拟通过扫描电镜高频振动夹持器的设计和轨迹规划的研究，得到磨粒SECM-SICM图像的快速采集方法，大量地收集磨粒剥落面上三维形貌和电位分布、化学成份特性图像，为故障诊断提供必需的数据集；并拟采用多目标优化的故障诊断方法，通过特征空间、多优化目标和算法结构的研究，实现故障诊断中复杂映射关系的解耦，最终完成基于磨粒SECM-SICM图像的发动机故障诊断方法，提高发动机故障监测与诊断的正确性和准确性。

磨粒携带着发动机重要的状态信息，对磨粒特征的获取是对发动机故障诊断的一个重要手段。然而受到检测技术的限制，对磨粒的检测只利用了其大小、二维形状等，未涉及更微观层次的结构和电化学分布的相关信息。这使得对发动机故障产生机制的认识不能做进一步深入的理解和研究。为了解决这一问题，进而有效地揭示发动机的工作状态和故障成因，本项目将扫描电化学显微镜（SECM）和扫描离子电导显微镜（SICM）相结合，采用一套全新的SECM-SICM检测系统实现对磨粒表面微观形貌和电化学信息的收集。在项目进行过程中，我们首先自主研发了一套SECM-SICM检测系统。其次，通过采用高频振动夹持器的设计、扫描轨迹规划算法、探针控制策略等角度对其成像稳定性和快速性的研究和改进，使其进一步满足了磨粒SECM-SICM图像的可靠高效的采集方法。最后，我们根据收集到发动机磨粒三维形貌和电化学信息，通过特征空间、多目标优化算法的研究，实现故障诊断中复杂映射关系的解耦，完成了基于磨粒SECM-SICM图像的发动机故障诊断方法。这种磨粒表面微观复合信息的测量方法将进一步促进人们对发动机故障产生机理的认识和对相关设备的改进，无论对于人民生产生活，还是国防军事均具有重大意义。