Font Size: a A A

Application Of Bayesian Networks Into Fault Diagnosis Of Crane

Posted on:2012-11-29Degree:MasterType:Thesis
Country:ChinaCandidate:Y Q ZhongFull Text:PDF
GTID:2232330395485366Subject:Software engineering
Abstract/Summary:PDF Full Text Request
Nowadays, fault diagnosis of large cranes with powerful and complex structurehas become an important research. For coup relationship and uncertainty informationexits within large parts of crane, the uncertainty of failure is always of uncertainty.Solving the uncertain diagnosis is the most important issue.Based on the study of fault diagnosis of large cranes and Bayesian networktheory, this paper propose a Bayesian network diagnostic model for fault diagnosis ofcranes. Taking the fault diagnosis system of a large crane as application object, weapply Bayesian diagnostic model to the system. Based on original rule-baseddiagnostic model using libraries. We add a method of using Bayesian network, variousforms of diagnostic reasoning can be done by new system and uncertainty issue forlarge cranes can be solved.There can three steps in Bayesian network model applyingto big cranes: first, construct libraries of Bayesian network used in fault diagnosis oflarge cranes. The structure and parameters of Bayesian network can be achieved byreasoning and probability scale method, after that, the Bayesian knowledge base isconstructed. Second is learning Bayesian network parameters. Maximum likelihoodestimation method is used to study and adjust Bayesian network parameters; third,construct the model of reasoning mechanism. Quantitative inference of fault can bedone by join tree algorithm. In the realization of context, the overall method ofdiagnosis system for Bayesian network is designed. Taking model as basic units,network generating module, learning modules, reasoning model and Bayesianknowledge model is designed, which laid a solid foundation for the softwaredevelopment of the entire crane system.Compare the original system model with the enhanced system with Bayesiandiagnosis, a better speed and accurate locate is achieved. The origin system has got anextension and improvements, which verify the effectiveness and application value offault diagnosis model and algorithm proposed by this paper.
Keywords/Search Tags:Fault diagnosis of cranes, Bayesian Networks, CPT, Probability scale, MLE, J1939
PDF Full Text Request
Related items