predictive control, measurable disturbances, quality of control, CARIMA model

Most of the processes that need to be controlled are affected by external disturbances. In many cases, the disturbances can be measured. A control method that can successfully cope with disturbances which are known in advance is predictive control. The disturbances known in advance can be incorporated into calculation of the manipulated variable in predictive control. In many control applications, the disturbances are measurable but unknown in advance. In this case, the future course of the disturbance must be estimated on the basis of its past values. This paper proposes different approaches to estimation of future course of the measurable disturbances in predictive control. Particular approaches were compared by simulation.