Fault Diagnosis Method For Replacement Process Based On Dynamic Causality Diagram

With the growth of social economy and the advancement of science and technology,the demand for high-quality metal products in various industries is increasing,while the high-quality mineral resources in China continue to decrease.Hydrometallurgy is widely used because it can effectively treat low-grade complex mineral raw materials and has low environmental pollution.The production scale of complex industrial processes represented by hydrometallurgy is becoming larger and larger,the production process and equipment are becoming more and more complicated,and the production process is continuous operation.Therefore,when there is an abnormality in the production process,it may cause great losses.Therefore,it is of great significance to diagnose the hydrometallurgical process.Since the hydrometallurgical process is a production process with complex processes,the process contains a large amount of qualitative information,quantitative information,and information and uncertainty information,the effect of fault diagnosis only based on data or mechanism model is often not ideal.As a knowledge-based fault diagnosis method,Dynamic Causality Diagram(DCD)can make up for the shortcomings of traditional fault diagnosis methods,so it has attracted extensive attention from experts and scholars.Based on the hydrometallurgical replacement process as the research background,and the replacement process mechanism is deeply analyzed,and its common process faults and deep fault causes are explored to realize the diagnosis of common faults in the replacement process.The fault diagnosis method based on dynamic causal map is mainly divided into two steps:the establishment of fault diagnosis model and fault diagnosis and reasoning.Firstly,the process mechanism and common faults of hydrometallurgical replacement process are analyzed.The mechanism knowledge and the causal graph structure learning method based on genetic algorithm are used to establish the fault diagnosis causal map of the replacement process.The parameters of the model are obtained according to the causal graph parameter learning method.Among them,the causal map parameter learning method mainly integrates the historical probability data of the historical data with the expert knowledge based on the DS evidence theory,and solves the inaccuracy caused by the uncertainty of the information and the use of historical data or expert knowledge to obtain the parameters.In order to further improve the accuracy of causality diagram fault diagnosis,a causal diagram reasoning method based on cloud model is proposed in this paper.The probability interval of the basic event and the connected event is obtained through parameter learning,the probability interval is converted into the form of cloud parameters,and the cloud parameters are substituted into the causality diagram inference calculation process according to the inference algorithm of the cloud model.The cloud model is used to express the inefficiencies in the replacement process.The simulation results of the conventional method and the improved method are compared to verify the advancedness and accuracy of the improved method.