基于先验简洁关联知识的故障诊断方法

Fault Diagnosis Method Based on Priori Concise Association Knowledge Between Fault States

在系统故障诊断优化问题的研究中,建立故障与故障表象之间的关系,对故障进行分析和预测是一种重要的故障检测手段。数据挖掘中的关联规则表达了事物间关系,在离散事件分析中具有分类和预测的功能。然而普通关联规则具有表达冗余量大和计算复杂的缺点,提出采用直接生成闭模式并生成的简洁关联规则的方法,挖掘并表达故障特征间的关联关系,并建立故障分析过程。在离散事件的分析系统中,改进方法相对于基于普通规则的处理方法,既减小了知识表达的空间,又提升了处理的效率。结果表明,提出的检测方法准确、有效,具有一定的实用性和推广价值。

Building relationship between faults and their expressions to diagnose fault is an effective method in a system. Association rule tells relationships between items. And it can function as classifier or predictor. But frequent item comprises too much duplicates, which can decrease mining efficiency. In this paper, a method was proposed based on concise association rules. It can prune search space with little cost and mine concise itemsets and rules di- rectly. Then fault diagnosis process was conducted through concise rules based on closed itemsets. The mining result consumes little space while the mining efficiency increases. The application shows that this approach runs stably and accurately.