The Research On Zinc Coating Thickness Model In Continuous Hot-dip Galvanizing Process

With the development of the industries of automobile, house appliance, the market demand for galvanized steel coil witnessed a huge growth. One important technologic index to value the product’s quality is the thickness and homogeneity of its coating. The control standard of the coating’s thickness directly determines the quality, cost and market competitiveness of the galvanized steel coil products. The producing line of hot-dip galvanization is a nonlinear and multivariable process characterized with a large delay, where air knife plays a decisive role in the control of the Zinc film’s quality. At the meantime, lack of mature mathematical model, most galvanization producing line still relies on the traditional method to adjust manually the set value according to operator’s experience, which inevitably causes deviation on coating’s thickness. Therefore, establishing a mathematical model between the parameters of air knife and the thickness of Zinc coating is a key step during the hot-dip galvanization process.Taking regard to the real technologies in the continuous hot dip galvanizing line of a certain band steel, the thesis aims to establish the Zinc coating thickness model on the basis of the intensive analysis of the characteristics of the process. The main researches are summarized as follows:1. With the blowing hot galvanization during the production process of the hot-dip galvanization of steel coil, the study analyzes the key factors in the process of air-knife’s blowing that influence the Zinc coating’s thickness. Also, based upon the mechanism of hydromechanics and fluid mechanics, the thesis finally figures out an analytic model between Zinc coating’s thickness and steel coil’s surface pressure distribution.2. To solve the unpredictable problems of uneven surface pressure distribution of band steel, this research takes use of numerical simulation method by combining FLUENT and ICEM software.At the same time, using the method of least squares regression analysis, it establish a model between the steel strip’s surface pressure distribution and different parameters of the air knife. Furthermore, it integrates this model with analytic model to build a mechanic model between the Zinc coating’s thickness and the parameters of the air knife.3. In order to alleviate the impact upon the model’s output results from some unknown and under-considered factors existing in the process of the model building, the present research uses LS-SVM to set up a grey box model to make up the data, which makes a hybrid model. The simulation result shows that the hybrid model is a superior model with a higher accuracy.Thus it verifies the model effectiveness.4. To solve the problems of parameters picking’s being time consuming, blind and easy to fall into partial optimization of LS-SVM hybrid model, we take advantage of particle swarm optimization method to optimize this model’s parameters, and with a higher convergence speed and accuracy. The simulation results show the effectiveness of the optimization algorithm, and LS-SVM hybrid model is more accurate after PSO.Through comparative analysis, LS-SVM hybrid model take use of the analytic model and data model, and improve the accuracy of the model effectively. PSO- LS-SVM hybrid model overcomes the shortcomings of LS-SVM parameters selection, with faster convergence and higher accuracy. PSO-LS-SVM hybrid model is a better model. If we apply it to the actual production, it will greatly improve the accuracy of the Zinc coating thickness control and make energy conservation.