Switching model adaptive predictive control

The paper considers the results of an adaptive predictive control system implemented in a combustion furnace. The main goal is to optimize the temperature control replacing the existing PID. The furnace dynamic is strongly nonlinear, has an important delay and its behavior depends on the operating point. These characteristics make the system very attractive for implementing a controller that achieves a more precise temperature control. The application consists of a weighted predictive control with its parameters obtained indirectly from a recursive identification of the plant. The project was developed at Siderca - Argentina, a seamless steel pipe factory. Copyright (C) 1998 IFAC.