Research On Modeling And Sliding Mode Control Of Cement Pre-Grinding System

As an important link in cement grinding,the core of cement pre-grinding system is to complete the preliminary grinding and powder selection of materials before entering the mill,which will help reduce energy consumption and improve the working efficiency of the mill.However,the cement pre-grinding system has the characteristics of many variables and complex process.At present,it mainly relies on knowledge workers to maintain the balance of cement pre-grinding system manually.Therefore,in order to realize the automatic control of cement pre-grinding system,it is of practical engineering value to combine advanced control methods and realize automatic control of cement pre-grinding system.Based on this,this paper carried out research on modeling and control algorithm of cement pre-grinding,with the main contents as follows:(1)Based on the analysis of cement pre-grinding process,the cement pre-grinding process is divided into steady flow warehouse weight control loop and mill load control loop.The variables that influence the stable flow bin weight and mill load are: main circulation fan speed,feed volume,current of separator and current of outlet bucket.After comparison of the field data,it is determined that the feeding amount and the speed of the main circulation fan are the control amount,and the steady flow warehouse weight and the lifting current of the mill hopper are the controlled amount.(2)In view of the difficulties in mechanism modeling and complex system process of cement pre-grinding system,this paper uses the data modeling method to establish the dynamic model of water pre-grinding system.The least squares and ant colony algorithm were used to build the feed-steady flow warehouse weight dynamic model.In order to establish the dynamic model of main circulating fan rotational speed and outlet bucket lifting current,genetic recursive least squares algorithm and deep circulating neural network were used respectively.Based on the analysis and comparison of modeling errors based on three modeling accuracy indexes,the dynamic model of feed volume-steady flow warehouse weight and the dynamic model of main circulating fan rotational speed-outlet bucket lifting current were obtained with high modeling accuracy.(3)Tracking problem of steady flow warehouse weight in cement pre-grinding system.The double power combination function reaching law and terminal sliding mode surface are designed in the stable current warehouse re-nominal system.The simulation experiment verifies that the sliding mode controller has better dynamic performance,which shows that the method has good performance in the fast convergence of stable current warehouse re-nominal system.In the disturbed stable current warehouse heavy system,combined with the disturbance compensation algorithm,a time-delay estimation based sliding mode controller is designed.The simulation verifies the anti-interference and fast convergence of the time-delay estimation based sliding mode controller,and realizes the fast and stable convergence of the stable current warehouse heavy in the disturbed state.(4)Aiming at the problem of mill load tracking in cement pre-grinding system.In the nominal system of mill load,the attenuating reaching law and global fast terminal sliding mode surface are designed.Compared with PID controller in simulation experiment,the sliding mode controller has better dynamic performance,and the fast and stable convergence of mill load in the nominal system is realized.In the disturbed mill load system,a sliding mode controller based on nonlinear disturbance observer is designed and compared with PID controller.The simulation results show that the disturbance observer can accurately estimate the fast convergence of the disturbance and the sliding mode controller and realize the fast and stable convergence of the mill load under the disturbed state.