Research On FMS Modeling And Fault Diagnosis Based On Petri Net

Flexible manufacturing system (FMS) is a high flexibility, high efficiency manufacturing system for small, moderate and large quantities of production, which has been concerned widely in the research of theoretical and engineering field. The flexible manufacturing technology maybe speed up the production process of new products, reduce the costs of production, improve the quality of production, increase the flexibility of production, increase the resilience of the market, and then obtain better economic benefit. So the research of the key technologies in FMS becomes an important topic. Modeling, fault diagnosis and reliability analysis are the key points to determine whether FMS can fit the practical application. The level of these three research points has become an important indication to evaluate the level of the whole FMS. As a formal and graphical analysis tools, Petri net is used in this paper for modeling, fault diagnosis and reliability analysis in FMS. The main research works and results as follows:(1) For the complex structure of FMS, a hierarchical modeling approach named UML-OOPN is proposed. The specific rules and algorithms of the UML model to OOPN model is given, using non-variable analysis method to complete the deadlock analysis of the system. The advantage of both UML and Petri has been played. The UML-OOPN can support all the needs of FMS from requirements analysis to the model realization in the full life cycle.(2) As there are no time constraints, type constraints in object-oriented Petri net model, so we can't carry out quantitative analysis for system. A new object-oriented extended colored timed Petri net is proposed in FMS to describe the time characteristics of the system, and analyze the accessibility of the model by the correlation matrix. The quantitative analysis of time characteristics is completed and the modeling method is verified.(3) In order to solve the uncertain of threshold and confidence degree in fuzzy production rules, a new self-learning fuzzy Petri net is proposed. By the processing to knowledge model hierarchically, we use neural network learning ability to adjust the weight value, threshold and confidence degree of the knowledge model at the same time. So the system has the ability of self-learning, which can learn new expertise's knowledge to do fault diagnosis autonomously. Fault diagnosis of FMS process based on self-learning fuzzy Petri nets are given, reflects the parallel inference that this method can show the advantages of fault dynamic spread.(4) In order to solve the possible state space explosion problem when reliability analysis work is done in a complex and large-scale system with stochastic Petri net, a hierarchical stochastic Petri net is proposed. A FMS reliability analysis model has been built to get the indicators of reliability and the system reliability index system is given. The software HSPN-Tool has been developed for the reliability of simulation and the system reliability of the simulation analysis is given by hierarchical stochastic Petri net. Through simulation, the key factors affecting system reliability is analyzed.