Optimal control of a steel strip rinsing process

Deals with optimal control of a steel strip rinsing process. The rinsing process is a dynamic nonlinear process. Modelling and identification of the process is based on a priori knowledge about the process and measured data from the process. Some parts of the process wear out. In the model, the worn parts are modelled explicitly and estimated on-line by an extended Kalman filter. The process is influenced by changing production variables, which are measurable but not controllable. The process is also influenced by disturbances. The optimal controller is based on a mathematical model of the process which includes on-line estimation of unknown parameters. The model is expressed in a discrete state space form, which makes the model suitable for optimal control. The physical limits of the process consist of the limits of the control signal. The optimal control signal is achieved by minimization of a quadratic loss function.<>