基于改进HMM的潜在电子故障状态识别模型

针对复杂电子装备隐性故障难以诊断的难题,在深入分析隐马尔可夫模型的核心问题及基本算法的基础上,探讨了其在故障诊断应用中存在的主要问题,建立了多状态电子装备可靠性评估模型,利用系统可靠性评估结果作为隐马尔可夫模型的初始模型特征量,改进了传统的隐马尔可夫模型,并对Baum-Welch训练算法进行了优化,形成了一套适于复杂电子装备潜在故障状态跟踪识别的数学模型.实验结果显示,理论方法及模型能够更好地识别潜在故障状态,加快了模型训练速度,提高了故障状态识别率.

Potential fault recognition based on improved hidden Markov model

Hidden failures are difficult to diagnose in complex electronic equipment.Based on in-depth analysis of the core issues and basic algorithm of hidden Markov model,the main problems of the algorithm in fault diagnosis application is studied,a reliable multi-state electronic equipment assessment model is established,the system reliability assessment result is used as the initial model features of hidden Markov model;and traditional hidden Markov model is improved,and the Baum-Welch training algorithm is optim...