Process Monitoring Method Of Reheating Furnace Based On Nonlinear Multivariate Statistical Theory

With the rapid development of science and technology and continuous progress,the modern production process industry has obtained rapid development also.There are usally many process variables and the correlation of the process variables are complex in the process of current industrial processes.Once the fault occurs,it will affect the normal production,which will cause huge losses of property and casualties.Therefore,the safety and reliability of the complex industrial system process has become one of the important issues,which makes the importance of monitoring and fault diagnosis of equipment condition more prominent.Process monitoring and diagnosing method based on multivariate statistical method have made great progress in theory and practical application nowadays.The content of this paper is derived from the multivariate statistical method for the on-line monitoring of operation parameters for reheating furnace process to observe the changes of these state parameters.Through the operation status parameters of the heating furnace production process real-time collection,and multivariate statistical analysis.As a basis of improving the online mathematical model of equipment.In a word,it is process monitoring method for the operation evaluation.Accurate monitoring of the operation process plays an important role for the eatablishing of the mathematical models in the process of reheating furnace.In this thesis,the process monitoring methods based on nonlinear multivariate statistics theory have been used.Through the analysis of the operating parameters of the heating furnace to verificating the mathematical model of the furnace operation process.Which provides a basis for the on-line calibration of total heat exchange factor.The main research is as follows:(1)The mathematical model of the reheating furnace based on total heat exchange factor and the influence of the mathematical model parameters are analyzed.And then the relationship between operation and total heat exchange factor had been determined.(2)Process monitoring methods based on kernel principal component analysis(KPCA)and kernel independent component analysis(KICA)are introduced.The two methods regard the reheating furnace process as the Gauss distribution and non Gauss distribution and then make experiment research and comparative analysis for the operating parameter of the heating zone.Two classes of slabs results prove the effectiveness of the methods.(3)Process monitoring methods based on renyi entropy of kernel entropy component analysis(KECA)is introduced here.Reheating furnace process is considered as a dynamical process here.Therefore,we apply KECA to the collected data.The experimental results show that the method is superior to KPCA and KICA methods.(4)According to the operation monitoring results,we make online calibration for the continuous production process of abnormal slabs.As a result,the mathematical models which have been calibrated have more accurate results compared to the original mathematical models.Therefore,the monitoring methods of this paper have a certain practical significance in practical production.