Research On Bayesian Network Algorithm For Fault Diagnosis System Of Cement Rotary Kiln

The cement rotary kiln is one of the most important equipment in the process of producing cement. Calcination is an important process in cement production which occurred in cement rotary kiln. The working state of the cement rotary kiln will directly affect the quality and yield of cement clinker. No matter how high the reliability of equipment, advanced production technology, failure is inevitable. The fault performance of cement rotary kiln is uncertainty and diversity as large size and complicated system. The traditional method of artificial investigation gradually become difficult to meet the requirements for the fault diagnosis of rotary kiln.Therefore, the production process and working principle based on the construction of the cement rotary kiln, cement kiln fault diagnosis model, and carries on the parameter learning and inference, realize the fault diagnosis of cement rotary kiln is accurate and efficient. This paper bases on the production process and working principle of the cement rotary kiln, structures cement kiln fault diagnosis model and carries on the parameter learning and inference. Then the accurate and efficient fault diagnosis of cement rotary kiln is realized.The specific contents are as follows:First of all, the paper introduces the relevant theoretical knowledge of Bayesian network, which is introduced by the structure learning, parameter learning and inference. In addition, this chapter mainly introduces some important classical algorithm.Secondly, two improved Bayesian network structure learning algorithms are proposed, which aim at the defects existing in the existing classical algorithm of Bias network structure. The performance of the two algorithms is verified and their advantages and disadvantages are analyzed by simulation experiment.Thirdly, study the process and mechanism of cement rotary kiln and analyze the property of the variables and the interaction between them. After enough collected,these datasets must be screened and quantization processed. In order to solve theproblem of fault diagnosis of the cement rotary kiln, Bayesian network is applied. We select the SHC algorithm to train datasets to establish the structure of Bayesian Network for the fault diagnosis of the cement rotary kiln.Finally, the Bayesian network parameter learning and inference process are introduced. The classical maximum likelihood Estimated method and variable elimination method are applied to study a Bayesian network model for fault diagnosis of cement rotary kiln in order to parameter learning and reasoning.