基于广义模糊熵的液体火箭发动机故障检测研究

定义了系统状态的广义模糊熵及其计算公式，即当前时刻滑动数据窗口上聚类结果的我模糊度，给出了滑动数据窗口中采样数据矢量方向分布中心的离散度定义及其与广义模糊熵之间的负指数关系，并基于该离散度概念对自适应窗口滤波器进行了改进，提高了滤波器对噪声对敏感度，从而进一步提高了基于自适应窗口滤波器的故障检测算法对强干扰噪声的鲁棒性。本文基于受强噪声干扰的实际试车数据，采用模糊Ｋｏｈｏｎｅｎ聚类网络（ＦＫＣＮ）

GENERAL FUZZY ENTROPY BASED-FAULT DETECTION AND ITS APPLICATION IN LIQUID ROCKET ENGINE

This paper presents a definition of general fuzzy entropy for multivariate data distribution,based on this definition,a measure of the distribution center divergence of the multivariate data within a sliding window is also given in this paper for the first time.The divergence definition provides an objective measure of the geometrical distribution divergence for the data vectors in a data window,it can be used as an ideal and objective criterion whether there exist heavily noised data in the current sliding window for the adaptive windowing filter,which was specially designed for fault detection and diagnosis before.Thus the noises discrimination ability is much improved,compared with the crisp and hard criterion previously used.This greatly improves the robustness of the fault detection algorithm to the heavy noises and sensitivity to subtle fault transients.In addition,the practical testing data with heavy noises was used to verify its efficiency and objectivity in fault detection and noises discrimination.