Supervision of a steel strip rinsing process

Condition supervision of steel strip rinsing process is considered. The rinsing process is a dynamic nonlinear process. Modeling and identification of the process is based on knowledge about the process and measured data from the process, known as gray-box identification. Some parts of the process wear out and are changed after manual inspection. In the model, the worn parts are modeled explicitly and estimated from measured data from the process. Data are collected before and after the worn parts are exchanged. The estimation is first made by offline identification by optimizing the likelihood function. Second, the estimation is made by using an extended Kalman filter. The result of the estimation is used to give a basis for a decision on which worn parts are to be exchanged