Modeling of reheating-furnace dynamics using neural network based on improved sequential-learning algorithm

In order to model the dynamics of a billet reheating furnace, a multi-input multi-output radial-basis-function neural network is constructed based on an improved sequential-learning algorithm. The algorithm employs an improved growing-and-pruning algorithm based on the concept of the significance of hidden neurons, and an extended Kalman filter improves the learning accuracy. Verification results show that the model thus obtained accurately predicts the temperatures of the various zones of the furnace