论文部分内容阅读
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。
Mechanical vibration has adverse effects on the working performance and the service life of large rotating machinery, and the fault diagnosis work is especially important. The traditional fault diagnosis based on the CP (Confidence-Probability) formula mainly relies on the subjective Bayes inference method to obtain the posterior probability of the fault source under the current observation. The limitation of this method lies in that it relies too much on the prior probability of the fault source. However, under normal circumstances , The prior probability of the fault source is difficult to have an accurate figure. In this paper, an improved CP formula is proposed, which uses the relationship between rule credibility and probability, replaces the prior probability of conclusion with rule credibility, adequacy measure and necessity measure, Conclusion The dependence of prior probability. Finally, a rotor test bed is used as an experimental device to simulate vibration faults and establish a knowledge base of mechanical vibration fault diagnosis. Experiments show that this method can be a better diagnosis of mechanical vibration fault diagnosis, with good practicality.