基于规则获取的Petri网故障诊断模型研究

On Petri Net Models for Fault Diagnosis Based on Rule Generation

针对构建合理的Petri网故障诊断模型,研究了以矩阵运算方式描述及处理故障发生过程中的逻辑关系的问题。以旋转机械故障诊断为例,用粗糙集理论对其样本条件属性进行约简,利用Skowron缺省规则获取方法得到故障诊断规则;并对传统Petri网进行改进并构造诊断模型,利用Petri网的动态描述与矩阵运算能力进行故障诊断,从而在一定程度上解决了样本集中存在不一致信息的问题,避免了Petri网对于复杂系统建模时出现的组合爆炸,具有一定的创新性和较大的使用价值。应用实例验证了该方法的可行性与有效性。

In accordance with establishing a logical Petri net fault model, the problem of handling logical relationship in fault occuring process through using matrix operating is studied. Taking the rotating machinery fault diagnosis as an example, the swatch condition attribute is redueod based on the rough set theory, and the diagnosis rule is got by Skowron default manner. A Petri net model is established to diagnose the fault state by using of matrix calculus, so as to avoid the matter that combinatorial states expanding quickly because of the complexity of the system and the matter of information being inconsistent. The experimental result shows the feasibility and efficiency of the proposed method.