基于磨粒分析方法的发动机磨损故障智能诊断技术

磨粒是研究磨损状态时最直接、最重要的信息元，通过对滑油中的磨粒进行监测与分析来判断机械设备的磨损情况，可以预防并监测机械设备的磨损故障。本文运用显微形态学方法建立了一套磨粒显微特征描述体系，以提取磨粒信息并进行磨损故障的模式识别；并结合摩擦学理论和人工智能方法，实现对发动机磨损故障的智能诊断和预测。

Intelligent Diagnostics for Engine Wear Failure Based on Debris Analysis

When the wear occurs, debris is the most direct and imp ortant information unit in the study of wearing process and fault mode. As the c haracteristics of wear particles have a close relationships with wear mode, fric tion mechanism and machine parts, it is foundation to uncover the friction mecha ni sm and the tribology character of contact parts. Through analyzing micrology sha pe of debris, this paper constructs a describing system of fractal parameters a nd identifies them effectively. With the characteristic parameters of debris, such as gray degree, fringe, texture, features of the figures, phase and so on, aided by artificial intelligence method (neural network), the method in which w ear particles are identified, classified and analyzed automatically, is presented. In addition, an actual monitoring experiment of some milit ary aero engine indicates that this intelligent diagnostic based on debris analy sis is effective to monitor machine condition and predict its wear failure.