Dead time and processes with stochastic disturbances

To control processes with unknown or variable dead time, it is important to estimate the dead time in real time. In the present work, adaptive control systems with reference models, where the estimation of the dead time in the batch processing model is considered, are studied. It is shown that the analysis of the output sensitivity variances with regard to the process dead time can allow a good estimation of this dead time. To estimate the dead time, it seems reasonable to start with a reference model corresponding to a high value of estimated dead time. This dead time estimation method was applied in a factory to a real process with significant stationary disturbances, and excellent results were obtained.<>