为了简化知识表示,并解决传统Petri网因样本有限和不确定而无法正确判定的问题,提出一种将模糊理论中的概率引入Petri网的理论.在HHT变换的基础上对故障数据进行特征提取,基于Rough集理论对所提取的属性特征进行属性约简,得出故障检测规则；将概率引入Petri网后,对约简后的规则建立基于概率Petri网的故障检测推理机制,实施在轨故障检测,并以地面轴承振动数据为例,对该方法进行验证；验证结果表明,该方法使用效果良好.%In order to predigest the denotation of the knowledge, and solve the problem that traditional Petri net can't make the right determinant because of the finity and uncertainty of the stylebook, a theory in which the probability of Blurry Theory is introduced to the Petri net is proposed. The fault characteristics of failure data is picked up based on the Hibert-Huang transition, and Rough set theory is used to reduce the characteristic attributes, then the fault detection rule can be concluded. Probability is introduced to the Petri net, and the probability Petri net reasoning model is built from the reduction rule to bring the On- orbit fault detection into effect. Oscillation data of axletree on ground is used as an example to validate the method, result shows that this method performs well.
展开▼
