Comparative research on prediction model of billet temperature for 1700 reheating furnace in Tangsteel company

In this paper, two modeling methods are studied for billet temperature prediction model based on the actual production data of the reheating furnace in Tangsteel 1700 line. The two methods are the neural network and the support vector machine. Two prediction models are built by the two methods respectively. And the comparative research is done via MATLAB simulation aiming at the two modeling methods. The research results show that the support vector machine method is more applicable when there are less sample data, but the neural network modeling method has the better accuracy than the other one when the sample data are enough.