基于独立成分分析的状态监控方法

针对目前部分的工业系统存在系统模型或系统独立变量不确定的情况,提出了一种基于独立成分分析(Independent Component Analysis,ICA)的状态监控方法。该方法不需要有关系统模型和系统变量的先验知识,解决了系统模型或系统变量不确定的问题。运用独立成分分析法对系统的观测信号进行特征提取,通过监控分离得到的特征信号来实现对整个系统的状态进行监控。分别对一个离散系统和一个双容水箱液位系统进行仿真验证,仿真结果表明该方法能够有效地实现系统监控。

State Monitoring Based on Independent Component Analysis

To the problems that system model or system independent components are indeterminate in some industrial systems,a new state monitoring method based on independent component analysis is presented.The method does not need priori knowledge about system model and system variables,and it solves the uncertainty of system model and system variables.The method uses independent component analysis to extract features of the state information,and monitors the separated source signals to achieve the whole system state monitoring.In order to verify the effectiveness and feasibility of the method,the new state monitoring method is applied to a discrete system and a double-tank system.The simulation results show that the presented state monitoring method is feasible and better.