Billet Temperature Modeling And Process Simulation Of Reheating Furnace

The billets must be heated to a certain temperature in hot mill production so that they can have a plasticity to be rolled. The quality of heating directly influences strip quality, products quantity, energy consumption and the life of mill equipments. Steel rolling heating-furnace control system is a multivariable and nonlinear system with large net lag and a high degree of coupling. Usually it is very difficult to measure the temperature of billets in real time so that most control system of heating furnace can only control the billet temperature by controlling the furnace temperature, thus the research of steel temperature model for heating furnace produce is of great significance.For a modeling process, the high quality of input data is a premise to achieve good effective model. Feature extraction for data recently attracted considerable attention, most of feature extraction techniques are based on linear technique such as principal components analysis (PCA) and independent component analysis (ICA), and they respectively make the principal components uncorrelated and independent. However, the traditional methods have linear requirements so that they fail to separate the nonlinearly mixed-signal, and somebody use "kernel tricks" to mapping the input data into feature space with high dimension where we make feature extraction with PCA and ICA algorithm that are so-called KPCA, KICA feature extraction method. LSSVM shows a good generalization and learning performance, and the good features obtained from feature extraction can be used as the modeling data of the LSSVM.. In this thesis, we predict the steel temperature with above method and the results show that the LSSVM has a high precision and can meet the request of engineering accuracy.Process simulation can replace the experiment that is difficult to implement. Not only it is visual, but also it can cut down the investment and reduce the risk. In this thesis, we use GENESIS 64 which is latest generation of 3-D automation configuration software to simulate the furnace process. First we store the corresponding variables needed in modeling, and the Matlab accesses the database with ODBC technique, then we predict the temperature of discharged billet by known model and input samples. We can predict the billet temperature online and achieve the simulation of reheating furnace process by the above flows.Finally, on the basis of review and prospect over the thesis, the future research directions are presented.