| Nowadays, Cement has become the most important construction materials in the world. However, China’s cement industry procedure at the same time also exist many problems, such as the lack of quality, high energy consumption, heavily pollution and so on. Therefore, in order to ensure cement production safety, it is much necessary to do some research on online monitoring and production quality control. Cement clinker quality determines the quality of cement productions, the content of free calcium oxide (f-CaO) in cement clinker can most directly, fastest to reflect the cement clinker quality. In order to improve the economic value of production, and ensure the cement clinker quality, it is important to analyze the parameter data which were collected from the online monitoring system.In actual production process, the cement production process is very complicated, online analyzing the content of f-CaO is very difficult, people usually use the glycerol-ethanol as extraction solvent, reaction with f-CaO generated calcium alkaline glycerol then use benzoic acid ethanol standard solution titration [8]. The operation is simple, easy to master, but low extraction rate of reaction caused by the long measurement time, at the same time extraction reaction generated H2O and is easy to cause the occurrence of other side effects, made the titration end point difficult to determine. It cannot satisfy the quality control requirement of modern large-scale cement clinker production. So the research on cement clinker free calcium oxide content of soft measurement modeling is very necessary.Soft measurement modeling is to construct a mathematical model, through the controlled variables and the disturbance factor, finding the function relation between the operating variables, to reflect the target variables which cannot be directly measured, so as to realize online prediction for the target variable. DCS control system in cement production line has been running a long time and accumulated a large amount of data, analysis of these data, find out the factors that influence the free calcium oxide content and prediction, this research can make a good foundation on f-CaO content monitoring on-line and cement clinker quality control, also can ensure production quality, improve the economic benefit and reduce the cost of production.This article mainly has carried on the following several aspects work:First of all, deeply introduce the new dry cement production process. Analysis the DCS control system by data mining technology and machine learning theory, researching the reason of f-CaO source and the impact on the quality of cement clinker, and determine the main influence factors of cement clinker f-CaO content is three rate values of raw material, rotary kiln coal feeding, decomposing furnace coal feeding, feeding rate, rotary kiln temperature; Secondly, based on the field operation parameter data, using the multivariate regression method such as least-squares method, principal component analysis and principal component regression, partial least squares, artificial neural network, such as:radial basis function, support vector machine (SVM) soft sensor intelligent control algorithm, establish the mathematical model of the cement clinker, the simulation results show that the established soft measurement model can effectively predict the content of f-CaO; Finally, comparison with the different the soft measurement model, determined the best soft measurement model. Comparison results show that the f-CaO content soft-measurement model based on support vector machine has the best prediction fuction. |
PDF Full Text Request