基于Petri网的顺序离散事件机电系统故障诊断方法的研究

发布时间：2018-01-26 09:03

本文关键词： 顺序离散事件机电系统故障诊断 Petri网 Bayes方法故障预测　出处：《上海交通大学》2011年硕士论文　论文类型：学位论文

【摘要】：随着现代科学技术的飞速发展,自动化装备的功能日益强大,结构日趋复杂,故障的发生也逐渐频繁,同时在许多大型复杂机电系统中存在着大量具有瞬时性,异步性,并带有顺序动作特点的离散事件机电系统。这类顺序离散事件机电系统有其自身的特点,但缺乏相关的研究手段,为了适应系统可靠性和稳定性的要求,有必要对这类顺序离散事件机电系统的故障诊断方法进行研究。研究归纳了顺序离散事件机电系统的主要类型,分析了顺序离散事件机电系统的故障特点,研究了其相应的建模方法。基于顺序离散事件机电系统的行为特点,分析了Petri网作为系统建模工具的依据,并在此基础上对三种主要类型建立基本Petri网模型,同时将时间统计信息引入Petri网,对顺序离散事件机电系统建立赋时Petri网模型。提出了基于Bayes试验方法的故障诊断算法,对变迁的时间统计量进行显著性检验,结合系统的先验和后验概率,在模块化故障诊断思想的指导下,分层定位至故障源。分析了在该算法下两类错误的概率,探讨了在最小误诊率的情况下,阈值概率的最优化问题,并推导了其计算方法。研究了微钻检测机的工作特点,分析了其作为顺序离散事件机电系统故障诊断实验平台的可行性。在故障诊断的基础上,研究了故障预测的方法,对顺序离散事件机电系统建立了模糊Petri网模型,推导了库所中托肯的模糊隶属度函数计算方法,提出了基于模糊推理的故障预测方法。在微钻检测设备上的实验结果表明,所提故障诊断和预测方法是可行的,具有较高的故障诊断准确率,本文所研究的相关内容可以为其他具有离散特征的工业故障诊断应用提供借鉴。
[Abstract]:With the rapid development of modern science and technology, the function of automation equipment is becoming more and more powerful, the structure is becoming more and more complex, and the faults occur frequently. At the same time, there are a large number of transient in many large complex electromechanical systems. Asynchronous discrete event electromechanical system with sequential action. This kind of sequential discrete event electromechanical system has its own characteristics, but lack of relevant research methods. In order to meet the requirements of system reliability and stability, it is necessary to study the fault diagnosis method of this kind of sequential discrete event electromechanical system. The main types of sequential discrete event electromechanical system are summarized and the fault characteristics of sequential discrete event electromechanical system are analyzed. Based on the behavior characteristics of sequential discrete event electromechanical system, the basis of Petri net as system modeling tool is analyzed. On this basis, the basic Petri net model is established for three main types, and the time statistics information is introduced into the Petri net at the same time. A timed Petri net model for sequential discrete event electromechanical systems is established. A fault diagnosis algorithm based on Bayes test method is proposed to test the significance of transition time statistics. Combined with the prior and posterior probability of the system, under the guidance of the modular fault diagnosis idea, the fault source is located in layers. The probability of two kinds of errors under the algorithm is analyzed, and the case of minimum misdiagnosis rate is discussed. The optimization problem of threshold probability and its calculation method are deduced, and the working characteristics of microdrill detector are studied. The feasibility of using it as an experimental platform for fault diagnosis of sequential discrete event electromechanical system is analyzed. On the basis of fault diagnosis, the method of fault prediction is studied. The fuzzy Petri net model for sequential discrete event electromechanical system is established. The calculation method of fuzzy membership function in the library is derived and the fault prediction method based on fuzzy reasoning is proposed. The experimental results on the microdrill detection equipment show that the proposed fault diagnosis and prediction method is feasible and has a higher fault diagnosis accuracy. The relevant contents of this paper can be used for reference for other industrial fault diagnosis applications with discrete characteristics.
【学位授予单位】：上海交通大学
【学位级别】：硕士
【学位授予年份】：2011
【分类号】：TH165.3

【参考文献】