A new kind of model of laminar cooling: By ls-svm and genetic algorithm

Abstract

In this paper, the conventional mechanism model of laminar flow cooling of hot-rolled strips is modified and optimized. And the new prediction model can express the laminar cooling process unambiguously, in which, the least squares support vector machines(LS-SVM) learning machine is introduced to promote the accuracy of the temperature-varying parameters and the genetic algorithm is proposed to identify the system of the temperature-varying parameters. The model can be used to calculate the temperature along the length direction and the thickness direction at the same time. The movement of the steel plate and the change of the valve can be obtained.