Industrial Products for Advanced Control of Mineral Processing Plants - Advanced Control and Supervision of Mineral Processing Plants - page 295

Figure 7.2 Main components of a fuzzy controller

function with given operating constraints and desired reference trajectories for pro-
cess outputs; and (3) repetition of the optimization process at each sampling instant
and application of the first value of the calculated control sequence (receding hori-
zon strategy). These three characteristics allow MPC to handle multivariable, non-
minimum phase, open-loop unstable and non-linear processes with a long time de-
lay and the inclusion, if necessary, of constraints for manipulated and/or controlled
variables. As shown in Figure 7.3, at instant t future process outputs are predicted
over the prediction horizon N p using an explicit model. The predictions depend on
known values of the manipulated and controlled variables and future control actions
u

( t + j | t ) over the control horizon N u . It is assumed that N u ≤ N p and that u ( t + j | t )

remains constant for j N p . The prediction horizon N p is related to the re-

sponse time of the process and N u is usually chosen so as to be equal to the order of
the model.

= N u , ...,

Figure 7.3 MPC strategy

The control sequence is obtained by optimizing an objective function that de-
scribes the goals the control strategy is intended to achieve. In classical MPC, an