摘要
提出不完备模型的两种不完备性:模型定义不完备和因果关系不完备.在模型定义不完备条件下,用在线观测与模型共同约束的方法处理观测乱序及未定义事件,得到可行的诊断轨迹.相对于基于完备模型假设下不能诊断的结论,该方法扩展了诊断方法的适用范围,放松了对模型的约束要求.在因果不完备条件下,提出用因果图联系部件,解决分布式诊断中由于部件独立建模而导致的不彻底诊断,提高了诊断的精确性.通过实验验证,两种条件下的诊断方法均能在相应的不完备模型中得到预期诊断结果,并对模型进行局部修订,提高模型的完备性.
There are two properties of incomplete： the incomplete of model definition and causality. With the condition of incomplete model definition, the method of constraint between online observation and the off-line model are proposed to process disordered and undefined events to obtain practical trajectory. Contrast to no beingdiagnosed by a complete model, this method expands the applicative scope and breaks the model limitation. On condition of causality incomplete, the usage of causal diagram to connect components is proposed. This methodsolves the halfway diagnostic problem caused by setting models separately, meanwhile enhancing accuracy. It has been tested that the diagnostic way under those two conditions brings out expected results according to certain incomplete models. It also reformulates model partially and improves the model maturity.
出处
《软件学报》
EI
CSCD
北大核心
2012年第3期465-475,共11页
Journal of Software
基金
国家自然科学基金(60973089
60873148
60773097
61003101)
吉林省科技发展计划项目基金(20101501
20100185
20090108
20080107)
教育部博士点专项基金(20100061110031)
浙江省自然科学基金(Y1100191)
欧盟合作项目(155776-EM-1-2009-1-IT-ERAMUNDUS-ECW-L12)
吉林大学符号计算与知识工程教育部重点实验室开放项目(93K-17-2009-K05)
