Predictive control of integrating processes based on a nonparametric model

The predictive control method presented was developed to assure a stable control of integrating processes that cannot be directly described with bounded weighting sequences. The predictive model is formed on the basis of a differential limited weighting sequence that is identified online. The length of the predictive horizon, is dependent on process characteristics. A delay, stability, and proportional or integrating process behavior are directly connected with weighting sequence member values. With the delay in a predictive algorithm the influence of nonminimal phase on the control stability was reduced. This means that processes with the nonminimal phase can also be controlled with a one-step predictor.