Research On Control System Of Tungsten Ion Exchange Smelting Process Based On CPS Principle

Tungsten is a kind of strategy element widely used in areas like metallurgy,aerospace and military.Tungsten ion exchange smelting process is a dynamic system with complex mechanism,many influencing factors and long process.Traditional control system which basically amied at the control of working procedure equipment is unable to solve the problem of coordinate control and parameter on-line prediction.Detailed investigation was conducted in production site of a tungsten smelting enterprise in Gannan,which is a listed company.Based on the analysis of smelting process features and the main influencing factors,and the principle of real-time perception and dynamic coordinate of CPS,a three-layer CPS control structure during smelting process was put forward.The main research work and conclusion are as follows:(1)The whole process of tungsten smelting was divided into several parts.The characteristics of the technological process was summarized.Then the controlled object and the input parameters were confirmed,by using partial correlation analysis method to find the main parameters that influence process indicators.Considering the robustness and stability of the system,a three-layer CPS control structure was put forward,which contained field layer,node control layer and decision layer.It provides a new solution for the control of tungsten smelting process.(2)According to the problem of insufficient adaptive capacity of local unit in master-slave control system,CPS intelligent node was designed based on real-time embedded technology,Through the connection of sensors equipment,sensing the status of production process in real-time,exchanging information with other nodes in the same layer,receiving optimized control strategy from decision layer.By adjusting the status of physical device through actuator,the problem of traditional control method local optimization and closed production were solved,the dynamic coordinate working effect was achieved.(3)In system decision layer,an on-line real-time prediction model was primarily studied.Historical data and process state parameters was input as model,using the excellent learning speed of ELM,through the constraints of ELM hidden layer weights and deviation by Bootstrap,its prediction accuracy was enhanced.Finally,the prediction of the terminal point of phosphorus removal in crude sodium tungstate solution during production process was conducted using the prediction model proposed in this paper.The result showed that the model had a good forecasting speed and accuracy,solving the problem of dynamic adjustment lag of the system had on operating parameters.Intelligent control node under structure and real-time prediction model were studied through stablishing the control structure of tungsten ion exchange smelting process using CPS principle.It is both practical and economical,for being able to reduce the fluctuation of product quality and process parameters,to improve the efficiency of production process.