Fault Diagnosis And Research Of Underflow Pipeline Of Thickener In Hydrometallurgy

Hydrometallurgy is a metallurgical process that treats complex ore and low grade ore and is less polluting to the environment.Solid-liquid separation plays an important role in the intermediate process of leaching and displacement in hydrometallurgy.In the dense washing process,the thickener is used as the main equipment for solid-liquid separation,and the equipment is expensive and complicated in structure.In the actual production site,the production environment of the thickener is rather harsh,and there are many interference factors,which make the thick washing process complicated,frequent faults,and even cause "compression","running" and other faults,so the wet metallurgy dense washing.The process is important for process monitoring and fault diagnosis.In this paper,the hydrothermal metallurgy dense washing process of a gold concentrator in Shandong Province is used as the background to diagnose the underflow pipeline of hydrometallurgy thickener.In this paper,the support vector machine(SVM)algorithm is used to establish the jamming fault detection model of the thickener underflow pipeline,and the clogging of the thickener underflow pipeline is detected.Support Vector Machine(SVM)is a machine learning algorithm that has advantages over previous machine learning methods in solving nonlinear,small sample and high dimensional pattern recognition problems.After the detection model is built,the particle swarm optimization(PSO)algorithm is used to optimize the parameters(nuclear parameters and error penalty parameters)of the established detection model to achieve better detection performance.The causes of the blockage of the bottom flow pipe of the thickener are mainly divided into three categories.In order to further diagnose the cause of the fault,this paper establishes a fault diagnosis model for the underflow pipe of the thick machine,diagnoses the detected blockage fault,and obtains the specific fault type so as to be targeted.Measures.By using the random forest algorithm to establish the fault diagnosis model,the random forest algorithm is a classic integrated learning algorithm,which can solve the problem that the ANN is easy to over-fitting and the convergence speed is too slow.In addition,the random forest can tolerate noise and abnormality well.Value,strong anti-interference ability.After the fault diagnosis model is built,the particle swarm optimization algorithm is used to optimize the parameters of the diagnostic model(the random forest decision tree and the number of attribute features randomly selected for generating a tree),and the satisfactory results are achieved.Finally,in order to better guide the monitoring of the thick washing process at the production site,the monitoring and fault diagnosis platform of the hydrometallurgy thickener was established by using Visual Studio C#and SQL Server database,and the fault diagnosis method was verified by relevant simulation analysis.Practicality.