Research On Control System Of Cement Raw Material Vertical Mill Based On Data Driven

Raw meal grinding is an indispensable part of cement production.The particle size of the finished product directly affects the quality of subsequent clinker firing,and it is the first "grinding" in "two grinding and one firing".As a kind of industrial process grinding equipment,vertical mill is widely used in the production of cement raw meal in recent years because of its high grinding efficiency,low power consumption and long life.The increase of equipment production capacity inevitably brings the problem of overcapacity.At the same time,most cement plants still use manual experience to adjust process parameters.Therefore,focusing on the key point of improving the automation level of cement raw material vertical grinding link,and on the premise of improving the quality and output,this paper puts forward a set of optimization control strategies including the upper set point optimization and the lower data-driven control loop,and gives the development process of the corresponding supporting control system.The specific research contents are as follows:(1)In order to achieve the desired goal of quality and output,a set value calculation method of internal pressure difference,material layer thickness and outlet temperature of the mill based on neural network in series is proposed,and a process of abnormal condition diagnosis and set value compensation based on rule-based reasoning and data mining is designed.Firstly,through the process analysis and expert experience of the raw material vertical grinding process,the dynamic relationship between the three underlying control loops was determined: the circulating fan speed-the internal pressure difference of the mill,the feeding amount-the material layer thickness,the cold air valve opening-the mill outlet temperature.Next,the upper set point calculation module is designed based on the method of neural network in series.The function of performance index is replaced by neural network,and the three set points are limited.Finally,taking the vibration amplitude of the mill as the judgment standard of abnormal conditions,the method of rule-based reasoning is used to judge the abnormal conditions,and the set value compensation is provided by data-driving method to output the final set value to the bottom control circuit.(2)In order to realize that the key parameters of the raw material vertical mill can accurately track the expected instructions,the design process of the raw material vertical mill controller based on data is given.For the sub loop of speed and pressure difference in the mill of the circulating fan,based on the data of speed and pressure difference in the mill of the circulating fan,an adaptive supervisor is designed by using the anti-counterfeiting algorithm.The PID controller is selected as the traditional controller,and the pseudo individual in the candidate controller is eliminated by calculating the performance index of the virtual reference signal to realize the adaptive switching of the PID controller.Aiming at the sub loop of feeding volume and layer thickness,using the information of real-time feeding volume and layer thickness,a PPD state observer of feeding volume and layer thickness is given to realize the dynamic relationship of feeding volume and layer thickness approaching online.On this basis,the controller is designed by FOPID method to realize the accurate tracking of the actual value of layer thickness to the set value.Aiming at the sub loop of cold air valve opening mill outlet temperature,using the real-time information of cold air valve opening and mill outlet temperature,the PPD state observer of cold air valve opening mill outlet temperature is given,which approaches the dynamic relationship of cold air valve opening mill outlet temperature on line.On this basis,the controller is designed by MFAC method,and the actual mill outlet temperature is realized by combining amplitude limiting and saturation compensation,realizes the accurate tracking of the actual value of the mill outlet temperature to the set value.(3)Based on the above calculation of set point and the research of data-driven control method,the software development process of automatic control system of cement raw material vertical mill based on data-driven is given.The design process of each functional module is described in detail,and through the interface design process,the specific operation specifications and the realization functions of the software are explained,which lays the foundation for the future field application.