Abnormal Condition Identification And Cause Tracing In The Process Of Hydrometallurgy Thickening

Hydrometallurgical process is the process of using liquid solvent to extract metal from the ore by leaching,solid-liquid separation and replacement.Thickener is the main equipment for solid-liquid separation in the hydrometallurgy process,the process of solid-liquid separation using a thickener is called "Thickening process".In practice,thickener operation environment is bad,there are many influencing factors,causing unstable operation frequently,it is easy to cause such problems as "hold scrappingslime rake" and "abnormal suspended solids".Once the fault occurs,not only the production stagnation,waste of raw materials,resulting in huge economic and property losses,and may even threaten the lives of field staff.The failure will not happen suddenly,the system will be in an unstable state before the failure,we call the state abnormal conditions,if the abnormal conditions can be effectively identified and accurately find the reason,you can avoid Fault occurred.The failure will not happen suddenly,the system will be in an unstable state before the failure,we call the system abnormal conditions,if the abnormal conditions can be effectively identified and accurately find the reason,you can avoid fault occurred.Therefore,it is of great significance to study the abnormal conditions of thickening process.In this thesis,aiming at the complicated working condition and the coupling relationship among many monitoring variables in thickening process of thickener,the identification and trace of abnormal working conditions are studied.The main research work is as follows:1)Starting from the working principle of thickener,the key variables in thickening washing process are analyzed,and the coupling relationship among many variables is analyzed.Starting from the typical fault state of the thickener,the abnormal state is deduced from the fault state,and the relationship between the abnormal state and each monitoring variable is analyzed.It lays the foundation for establishing a reasonable abnormal condition identification and cause tracing model.2)In view of the complex working condition and coupling relationship of thickening process,in order to combine expert knowledge and actual data analysis,the Bayesian network is used to study.For the missing and wrong data in the actual industrial production,the SEM algorithm under incomplete data set is chosen to study the structure of Bayesian networks.The algorithm is improved based on expert knowledge.knowledge-learning SEM algorithm is proposed,Introduce expert knowledge in the process of structure learning.Provide a more effective method for establishing Bayesian network model.3)Based on the knowledge-learning SEM algorithm and the data and expert knowledge,a Bayesian network model for identifying abnormal working conditions in dense washing process and its reasoning is established.According to the characteristics and properties of Bayesian network,the method based on the phenomenon layer node and the method based on incomplete information are used to study.Simulation results show that the Bayesian network based on the knowledge-learning SEM algorithm can effectively identify the abnormal working conditions and trace the reason of the thickening process.