Research On Automatic Control Of Cement Raw Material Quality Based On Typical Working Conditions

Cement raw meal proportioning is an important link in cement production.The quality of cement raw meal will directly affect the quality of clinker.There are many problems in raw meal proportioning,such as large lag,non-linearity and frequent fluctuation of working conditions.In domestic cement factories,fluorescence analyzer is often used to detect raw meal quality,which is controlled by manual operation.The precision of manual adjustment is limited,and the qualified rate of raw meal needs to be improved.The fluorescence analyzer detects raw meal every hour,and the detection cycle is long.In order to realize the real-time adjustment of cement meal proportioning,the project takes the 2500t/d cement production line of a cement plant as the research background,and on the basis of thorough understanding of cement raw meal proportioning technology,combined with the Swiss SpectraFlow company's near infrared online analyzer and advanced control algorithm,complete the scheme design and software development of automatic control system for cement raw meal quality.The main research contents are as follows:(1)Data processing and trend information extraction.Aiming at the problem of abnormal data and random errors in the process of data acquisition,firstly,the limiting filter method is used to eliminate the abnormal values,and then the weighted recursive average filtering method is used to smooth the data.The data filtering process is realized by combining the two methods.Aiming at the problem of trend information extraction,least squares polynomial fitting is used to extract trend information from data.(2)The division of typical working conditions and model establishment.Aiming at the frequent fluctuation of working conditions in cement raw meal proportioning,three typical working conditions are classified by using K-Means clustering algorithm,K-Medoids clustering algorithm and Fuzzy C-Means clustering algorithm under the condition that limestone proportion is not positively correlated with CaO content.Under three typical conditions,the ARX model was used to establish a four-in-four-out model with four raw material ratios as inputs and four oxide contents as outputs.(3)Controller design.Aiming at the large lag problem in the aw meal proportioning link,the model predictive controller is used to predict the output data of the future period by using the model established under each typical working condition,so as to realize the timely adjustment of the ratio and solve the lag problem.The laboratory fluorescence detection value is highly accurate,online analyzer detection cycle is short.According to the corresponding relationship between fluorescence detection value and online analyzer detection value,expert system control rules are summarized and expert system controller is compiled.(4)Design and development of automatic control software for cement raw meal quality.According to the above research contents,the automatic control system of cement raw meal quality is developed.The main developed modules include data filtering module,working condition identification and switching module,model prediction module and expert system control module.The field application shows that the software can run reliably for a long time in the field,and the control effect is good,which is helpful to improve the quality of raw meal and has strong practicability.