Fault Diagnosis Method Of Hydrometallurgy Process Based On Rough Set

Hydrometallurgy is a process for the extraction of ores by using liquid reagents to leach,separate,and extract metals and their compounds.Hydrometallurgy is widely used in metal smelting because of its advantages of being clean,efficient and suitable for the recovery of low-grade metal mineral resources.There are many complex influential factors and huge number of uncertain information in hydrometallurgy process.Many scholars have been studying the complicated mechanism of hydrometallurgy process.The problems that quantitative information coexist with qualitative information and deterministic information coexist with uncertain information must be solved when diagnosing faults in hydrometallurgy.The quantitative information refers to the variable information described by the numerical size,and the qualitative information refers to the variable information through qualitative description,which mainly includes semantics.Traditional methods of fault diagnosis,such as principal component analysis and other methods,can only deal with quantitative information.To solve the above problems,this dissertation takes fault diagnosis of gold smelting process as the research background,and presents a method of fault diagnosis of hydrometallurgy process based on rough set.This method is used in the fault diagnosis of the hydrometallurgy cyanidation leaching process and thickener decantation process.To deal with the coexistence situation of continuous and discrete attributes in the decision table,the ant colony clustering algorithm is used to discretize the continuous attributes in this paper.All continuous attributes are clustered together to form a comprehensive attribute in this method,which realizes the discretization of the continuous attributes.Compared with the traditional discretization method,it can better reflect the correlation and complementarity between different continuous attributes,and reduce the discretization of errors caused by the sample data on the boundary.By genetic algorithm,the attribute reduction is transformed into the evolutionary optimization problem in genetic algorithm by genetic algorithm,and then the attribute reduction is realized.Through rule acquisition,a fault diagnosis model of hydrometallurgy leaching process based on rough sets is established,and the simulation results verify the feasibility of the method.A fault tracking method based on contribution rate is also proposed in this paper.It can find the cause variables that cause the fault to realize self-healing control of production,and the feasibility of the method is verified by simulation.Aiming at the coexistence of deterministic information and uncertainty information in hydrometallurgy thickener decantation process,the interval data expression is introduced into the decision table,and the Hausdorff distance is introduced into the Interval clustering algorithm as the metric interval to be used in discretization of interval attributes.The discrete attributes and other discrete attributes constitute the rough set decision table.After attribute reduction and rule acquisition,a fault diagnosis mechanism model of hydrometallurgy thickener decantation process based on interval rough sets is established.The feasibility of the method is verified by simulation.