机械系统的故障演化测试性建模及预计

Fault evolution testability modeling and prediction for mechanical systems

采用部件级及系统级的分层建模思想,提出基于故障演化的测试性建模方法.在部件级,分析系统的故障演化机理,建立描述系统故障与征兆参数的定性相关性矩阵;在系统级,使用键合图的分析方法,建立系统中部件故障与子系统以及系统参数间的能量传递关系,进而获得征兆参数与测试间的动力学关系.在考虑故障可检测率、故障可隔离率的基础上,以故障可预测率以及故障可跟踪率分别描述系统对故障的检测、隔离、预测和跟踪的测试性水平.案例分析结果表明:故障与征兆参数的定性关系以及征兆参数与测试的动力学关系能准确刻画机械系统中故障演化与测试的关系,在此基础上提出的测试性预计方法能够有效评估机械系统对故障的检测、隔离、跟踪和预测能力,有助于提高机械系统对故障演化的跟踪和对失效的预测水平.

A modeling approach based on fault evolution testability model（FETM）was proposed by adopting a multi-level modeling scheme that combined system level and component-level models.The approach adopted a multi-level modeling scheme that combined system-level and component-level models.For component level,the fault evolution mechanism in systems was analyzed,then the qualitative dependency matrix was built to describe the relations between fault and symptom parameters.For system level,bond graph methodology was used to model the relations of energy transfer among faults that occurred in components,the parameters of subsystems and systems,thus the dynamic relations between symptom parameters and tests could be obtained.On the basis of fault detection rate and fault isolation rate,fault tracking rate and fault prediction rate were developed to describe the testability of detecting,isolating,tracking and predicting failures,respectively.The results of the case study show that the relations between fault evolution and tests can be described correctly by the qualitative relations between faults and symptom parameters,and the dynamic relations between symptom parameters and the tests of systems.According to those relations,the testability prediction can be utilized to evaluate the systems＇ capabilities of detecting,isolating,tracking and predicting failures.And the prediction results contribute to improve the ability of mechanical systems of tracking fault evolution process and predicting fault.