Data-driven Grinding Operation Optimization Control Method And Hardware-In-the-Loop Simulation System

In industrial production,the grinding process fully breaks ore to achieve single crystal dissociation between the mineral and the gangue,which is an indispensable part of the whole process of mineral processing.The operation indexes based on grinding particle size and cyclic load directly restrict the product quality and production efficiency.Because the influence of the basic control loop in the production process is lower than which the set-value control cause,the grinding operation optimization control method based on the set-point has been the focus of researchers for grinding process.Therefore,this paper carried out the research on the optimization control for grinding process under the support of the National Natural Science Foundation project “ Dynamic environment full-process data-driven operation optimization control”,using reinforcement learning and similarity-based transfer learning to design control method which use data to obtain the optimal set point.To verify the effectiveness of the above control method,hardware-in-the-loop simulation system for grinding process is designed and developed.The main research work of this paper is as follows:(1)For the problem that the optimization control method based on model or intelligence is difficult to realize the optimization control of the whole process grinding,Q-learning operation optimization control method based on neural network is proposed on this article,however,it is difficult to follow the change of working conditions for the learning speed of neural network.So this paper transfers the Q model under similar conditions to the current working conditions by case-based reasoning,accelerate the learning speed of the neural network,and achieve the goal of optimal control of the whole process of grinding(2)In order to verify the reliability of the operation optimization control method,a set of hardware-in-the-loop simulation system for grinding process was designed.Using METSIM and OPC technology to develop virtual objects in the grinding process,using the rapid control prototype technology to design and develop the operation optimization controller,and shorten the time of Simulink simulation to actual controller algorithm implementation.(3)In order to realize goals including adjusting parameters and observing operating state in the algorithm of controller,designing and developing algorithm monitoring software based on WPF,whose mainly features include experimental management,controller configuration,data display and variable management.(4)On the hardware-in-the-loop simulation system for grinding process,buid control model in Simulink to verify the effectiveness of the designed operation optimization method and to prove the reliability of the simulation system.