Design Of Wireless Soft-sensing System For Flotation Of Copper Ore

Flotation technology is the most important method of copper ore dressing and has been widely used.In the flotation process,the grade of copper concentrate determines the quality of the final product,so it is the key variable in the whole process.However,in the actual production,due to the highly complex and nonlinear flotation process,the measurement of this parameter is time-consuming and difficult to be detected online in real time.In addition,the existing on-site flotation process mostly adopts wired sensor network,which will bring a series of problems such as easy aging of communication lines and criss-crossing of wiring,which will not only affect the stability of communication system,but also increase the difficulty of flotation process control.To solve the above problems,this paper designs a monitoring system based on ZigBee wireless network and machine learning technology,which realizes the real-time soft sensor of copper concentrate grade in flotation process.The specific content is as follows:First of all,in view of the complex mechanism of flotation process and the large amount of redundancy among variables,a soft sensor modeling algorithm based on the combination of nonnegative garrote(NNG)and least squares support vector machine(LSSVM)is proposed.As a new nonlinear modeling tool,LSSVM can fully approximate arbitrary complex nonlinear relations;NNG algorithm has the ability to delete redundant variables and compress the coefficients of corresponding variables,which can reduce the correlation between variables.The effective combination of NNG and LSSVM can achieve the precise soft sensor of copper concentrate grade in flotation process.Secondly,on the basis of the study of soft measurement algorithm,the soft measurement model developed by MATLAB is imported into the development environment of IAR,The prediction model with good performance is trained by programming in C language,and the program is compiled and downloaded to the embedded system based on ZigBee,and the prediction model transplantation and online application based on MATLAB are realized on the designed hardware system.Thirdly,the PCB circuit board of the embedded system used by the lower computer is designed,including the core board of the embedded system and each functional module circuit,such as 485 communication circuit,serial port circuit,power supply circuit,level conversion circuit,step-down circuit,etc.,and the comprehensive debugging and performance testing of the hardware are also carried out.The experimental results show that the designed hardware has reached the PCB factory standard in terms of layout,wiring and system stability.Finally,in order to further achieve the copper concentrate grade value and the real-time monitoring of the field parameters,this paper also designs the upper computer monitoring program based on LabVIEW,and carried out the comprehensive test of the whole software and hardware system.The results show that the monitoring system can realize the real-time measurement and monitoring of parameters such as copper concentrate grade value,on-site temperature and humidity,etc.,and meet the requirements of enterprises for the measurement accuracy and real-time performance of key parameters.