Research On Sliding Mode Optimal Control Of Cement Combined Grinding Based On Data Driven

As the last link of cement preparation process,cement grinding is directly related to cement fineness and power consumption index.At present,most cement enterprises still rely on the experience of knowledge workers to control the cement grinding process,and there are some problems such as low working efficiency of roller press and unstable load of mill,which lead to high power consumption and unstable cement fineness.Therefore,considering the cement fineness and energy consumption index,focusing on stabilizing cement fineness and reducing power consumption,this paper puts forward the sliding mode optimization control strategy of cement combined grinding based on data drive and develops the control software suitable for cement grinding enterprises in China.The specific research contents are as follows:In order to ensure the comprehensive optimization of energy consumption ratio and cement fineness,considering the constraints of actual physical system,an optimization method of cement grinding index based on multi-objective dragonfly algorithm and rule-based reasoning is given.Firstly,according to the characteristics of cement combined grinding process,considering the power consumption and cement fineness as performance indicators,the weighing bin weight,mill current and cement particle size distribution are selected as key variables;Secondly,three key control links are determined: 1)feeding amount-stable flow bin weight,2)circulating fan speed-ball mill current,3)classifier speed-particle size distribution;Finally,under the limit of expert rules,multi-objective dragonfly algorithm is used to give the expected values of stable flow bin weight,ball mill current and particle size distribution,so as to achieve the best power consumption ratio and cement fineness during cement grinding.In order to realize the task of tracking the bin weight,ball mill current and particle size distribution,a data-driven sliding mode control strategy for cement grinding is proposed.Because of the complex grinding mechanism,it is difficult to establish an accurate dynamic relationship.In this paper,a data model is established based on data-driven theory by using production input and output data.Aiming at the control loop of feed rate-stable flow bin weight,based on its dynamic linearization data model,an integral sliding mode controller is designed by introducing dynamic constraint function to realize the stability of stable flow bin weight.Aiming at the current control loop of circulating fan-ball mill,a sliding mode control method of ball mill current terminal based on predetermined performance is given,in which,in order to ensure that the mill current is always within the dynamic constraint range,an error conversion function is introduced to convert the limited mill current tracking error into an unconstrained tracking error form to ensure that the mill current always runs in the preset area;Aiming at the speed-particle size distribution control loop of the classifier,a new error conversion function and a preset performance function are designed to limit the particle size distribution within the preset process range,and the design steps of the particle size distribution controller are given on this basis to achieve stable particle size distribution.Considering the cement fineness,power consumption and other multi-indexes,a set of multi-objective optimization control system software for cement combined grinding is developed,combining the optimization method of cement grinding indexes and controller design method based on multi-objective dragonfly algorithm and rule reasoning.Through field test,the self-regulation of cement grinding process can be realized,and the comparison of actual application data shows that the developed system can improve the qualified rate of cement fineness by 9.1% and reduce the power consumption of cement by 1.18%.