Research On Prediction Of Firing Zone Temperature In Roller Kiln Based On Mechanism And Data Hybrid Driven

Roller kiln is the core equipment of the ceramic production process,and its internal temperature distribution directly affects the quality and output of ceramic products,thereby affecting the energy consumption and cost of the entire enterprise.In the ceramic firing process,the firing zone has the greatest impact on the firing quality of ceramic products.The temperature adjustment inside the firing zone determines the quality of the product.Therefore,study the temperature and prediction methods of the roller kiln firing zone,and use the temperature prediction results to guide the operator to adjust the parameters of the equipment in advance,which can prevent the ceramic products from deforming,foaming,and matt quality problems,and effectively improve the operating efficiency of roller kilns,thereby realizing the energy saving and consumption reduction of the enterprise.According to the characteristics of strong nonlinearity,strong coupling,strong hysteresis and time variation in the production process of firing zone in roller kiln,a mechanism model of flue gas temperature in the process of heat transfer of firing zone in roller kiln was established based on the analysis of the production process and temperature influencing factors of firing zone in roller kiln.Then,an improved kernel principal component regression method based on data drive is proposed to build the temperature error compensation model.Finally,a hybrid model was obtained by combining the mechanism model with the error compensation model,which was used to predict the sintering zone temperature.The specific research contents of this paper are as follows:(1)Aiming at the firing process of ceramic products,the structural characteristics,temperature characteristics and firing system of firing zone in roller kiln were systematically analyzed.Then,according to the characteristics of the roller kiln,the key factors affecting the internal temperature of the roller kiln and the quality of ceramic products are analyzed,the temperature prediction framework is established,and the temperature prediction method suitable for firing zone of roller kiln is analyzed and the best algorithm is determined,which provides the basis for the selection of the temperature prediction method for the subsequent firing zone.(2)Aiming at the problem that the flue gas temperature in the sintering zone of roller kiln can not be measured accurately and in real time,which affects the quality of ceramic products,a one-dimensional steady-state mechanism model of the flue gas temperature in the sintering zone was established.Firstly,the heat transfer process of the sintering zone is analyzed,and the flue gas temperature distribution model is established according to the heat transfer process.Secondly,the solution process of the differential equation is analyzed and the method used to solve the mechanism model is determined.Finally,the simulation results are compared and the research content of the next chapter is introduced.(3)Aiming at the problem that the mechanism model is not accurate,a hybrid model based on mechanism and data drive is proposed to predict the sintering zone temperature.Firstly,an improved kernel principal component regression(KPCA)method was proposed to compensate the temperature error,and an error compensation model was established to solve the problem of the error between mechanism temperature and actual temperature.Secondly,the temperature error compensation model was combined with the mechanism model to form a hybrid model,and the hybrid model was used to predict the sintering zone temperature.Finally,the improved method is compared with the local weighted kernel principal component regression method and the moving window local weighted kernel principal component regression method.The results show that the improved method has better robustness and faster calculation speed,and meets the requirements of industrial production.(4)According to the specific requirements of the enterprise’s production,combined with the above research content,the integrated development and application of the temperature prediction module of the firing zone in roller kiln was realized by using the MATALB GUI software.The mechanism model and the temperature prediction model are integrated,and then based on the software integration method and Java language,the roller kiln firing zone temperature prediction system is implemented and applied to the actual production guidance.