【摘要】 机械振动对于大型旋转机械的工作性能和寿命均产生不利的影响,其故障诊断工作显得尤为重要。传统基于CP(Confidence-Probability)公式的故障诊断主要依据主观Bayes推理方法,得出当前观测下故障源的后验概率,该方法的局限性在于过分依托故障源的先验概率,然而通常情况下,故障源的先验概率很难有一个准确的数字。本文提出了一种改进的CP公式,该公式利用规则可信度与概率之间的关系,用规则的可信度、充分性度量以及必要性度量代替了结论的先验概率,大大减少了对结论先验概率的依赖程度。最后以转子试验台为实验装置,模拟振动故障并建立机械振动故障诊断知识库。实验表明,该方法可以较好的完成机械振动的故障诊断工作,具有良好的实用性。更多还原

【Abstract】 Mechanical vibration is adverse to working performance and life of rotating machinery.The fault diagnosis is important particularly.The traditional fault diagnosis based on CP formulas can get the posterior probability of failure sources under the current observations,but it depends on the priori probability of failure seriously.Generally,the accurate priori probability of failure is difficult to get.In this paper,considering the relation between the credibility and probability,an improved formula of CP was derived.The improved formula replaces the prior probability using credibility,adequacy measure and necessity measure.The credibility,adequacy measure and necessity measure can be easily given by experts and customs.Relations of credibility,adequacy measure and necessity measure also were discussed in the paper which help experts give more reasonable rules.Finally,the vibration faults were simulated using an vibration test equipment and knowledge database of faults was established.The results show that this method can complete the fault diagnosis of mechanical vibration and it has good usability.更多还原