Font Size: a A A

Research And Application Of Reliable Analysis Method For Complex System Based On Bayesian Network

Posted on:2017-04-23Degree:DoctorType:Dissertation
Country:ChinaCandidate:S ZhaiFull Text:PDF
GTID:1108330482497721Subject:Mechanical design and theory
Abstract/Summary:PDF Full Text Request
With the rapid development of modern industrial technology, the products and equipment systems have become increasingly complex. The complexity of the system, on the one hand, is embodied in its mutual coupling between subsystems or components; on the other hand, embodied in the variety of external influence factors such as the change of the work environment of system. And, due to the limitation of material, space and time, sufficient data and information cannot be obtained to make clear and accurate judgments on the state, characteristics and behaviors of the system, which leads to a lot of uncertainty in the system. The traditional reliability analysis method has obvious deficiencies and limitations in solving practical problems. The subjective and objective uncertainty in the reliability analysis of complex systems is dealt with in this dissertation, and the Bayesian network theory is used as the theoretical basis of uncertainty analysis. The fuzzy theory is combined with the traditional reliability analysis theory, with the battery production line being as the research object, the insufficiency of the present reliability analysis theory is analyzed, and the corresponding method and the analysis model is being built. The primary research contributions are summarized as follows:(1) The limitations of traditional reliability analysis method of fault tree and Bayesian network have been analyzed, and the reliability analysis method of fuzzy Bayesian network based on fault tree has been proposed. Bayesian network modeling method is applied to the basic modeling, and the event multi-state of the complex system is described by node multi-state expression features of Bayesian network theory, and the uncertainty of the logical relationship between complex system events is represented by node conditional probability table of the Bayesian network theory. In the framework of Bayesian network model, the fuzzy set theory is introduced to describe the expert fuzzy assessment of occurrence probability. In the process of gathering the expert evaluation information of uncertain weights, the method is proposed to integrate the experts’evaluation information of uncertain weights by relying on uncertain ordered weighted averaging operator in order to realize the objective weighting of the expert weight.(2) The limitation of the traditional FMECA analysis method has been analyzed, and the reliability analysis method based on FMECA fuzzy Bayesian networks has been presented. The fuzzy number of fuzzy theory is adopted to represent the expert’s fuzzy rating to RPN attribute parameter, and the traditional fuzzy rule base is replaced with the belief structure, which is used to describe the uncertainty between premises and conclusions of fuzzy rules under the condition of the incomplete fuzzy data input. The fuzzy rules are offered by using Bayesian network inference to integrate the belief structures to implement the fuzzy rules reasoning, and the detailed modeling methods and steps are proposed. The defuzzying method is put forward by using the weighted average to realize the explicit of fault damage levels.(3) The current problems existing in reliability analysis of complex system have been analyzed, and the application of methods combined with each other in reliability analysis is studied. The reliability analysis in battery production line system is taken as an example in this dissertation. Based on the reliability analysis in battery production line system under the condition of a variety of subjective and objective uncertain information, the integrated application of reliability analysis method is studied. In this case, the reliability analysis method based on FMECA fuzzy Bayesian network is proposed to determine the key important subsystem, and the reliability analysis method based on fault tree is adopted to determine the main components which influence the system and its failure modes, which can achieve the reliability analysis of the system qualitatively and quantitatively.
Keywords/Search Tags:Uncertainty, Reliability Analysis, Fuzzy Theory, Bayesian Networks, Fault Tree, FMECA
PDF Full Text Request
Related items