Page 227 - Distributed model predictive control for plant-wide systems
P. 227

Cooperative Distributed Predictive Control with Constraints            201


             Consider
                              √
                                ‖ f           ‖
                               2‖x (k + N − 1|k)‖ − ‖̂ x (k + N − 1|k − 1, i)‖ P
                                ‖             ‖P
                                      ‖ f          ‖
                                ≤ 0.42‖x (k + N − 1|k)‖
                                      ‖            ‖P
                                    ‖ f                             ‖
                                  + ‖x (k + N − 1|k) − ̂ x(k + N − 1|k − 1, i)‖
                                    ‖                               ‖P
                                ≤ 0.42  e +       e
             and substitute Equations (9.33)–(9.35) into Equation (9.32), it yields

                                V − V k−1,i
                                 k,i
                                                 ′
                                   ≤ −  e +(N − 1)         e + 0.42  e +       e
                                                          ′
                                   =−e(   −   (0.42 +((N − 1)   + 1)    ))
             which, in view of (9.28), implies that
                                            V − V     < 0
                                             k,i  k−1,i
               Thus, for any k ≥ 0, if x(k)∈ X∖Ω(  ), then there is a constant    ∈ (0, ∞) such that
                                                                       i         ∑ m
             V ≤ V     −    . Furthermore, we have the inequality of V ≤ V  −   , where    =    
              k,i  k − 1,i  i                                k   k − 1             i=1 i
             since m is limited. From this inequality, it follows by contradiction that there exists a finite
                  ′
                              ′
             time k such that x(k ) ∈Ω(  ). If this is not the case, the inequality implies V → −∞ as
                                                                              k
                                                               ′
                                                                           ′
             k → ∞. However, V ≥ 0, therefore, there exists a finite time k such that x(k ) ∈Ω(  ). This
                             k
             concludes the proof.
               So far, the feasibility and stability of proposed C-DMPC are analyzed. Provided an initially
             feasible solution could be found, subsequent feasibility of the algorithm is guaranteed at every
             update, and the resulting closed-loop system is asymptotically stable at the origin.
             9.5  Simulation
             The multizone building temperature regulation systems are a class of typical spatially dis-
             tributed systems, as shown in Figure 9.2, which are composed of many physically interacted
             subsystems (rooms or zones) labeled with S , S , …, respectively. The thermal influences
                                                  1  2
             between rooms of the same building occur through internal walls (the internal walls isolation
             is weak) and/or door openings. A thermal-meter and a heater (or air condition) are installed in
             each zone, which is used to measure and adjust the temperature of the multizone building.
               For simplicity, the seven-zones building is taken as the example. The relationship among
             these seven zones is also shown in Figure 9.2, where zone S is impacted by zone S and zone
                                                            1                  2
             S ; zone S is impacted by zone S , S , and zone S ; zone S is impacted by zone S , S , and
                                                      7
                     2
              7
                                                                                  4
                                                                               2
                                                             3
                                           3
                                        1
             zone S ; zone S is impacted by zone S , S , and zone S ; zone S is impacted by zone S ,
                                                                  5
                  7
                          4
                                                           7
                                                                                      4
                                                5
                                             3
             S , and zone S ; zone S is impacted by zone S and zone S ; zone S is impacted by all the
                                                             7
                                                                     4
              6
                                                   5
                         7
                                6
             other zones. Let U be defined to reflect both the constraint on the input u ∈ [u , u ] and the
                                                                              i,U
                                                                           i,L
                           i
                                                                      i
             constraint on the increment of the input Δu ∈[Δu , Δu ].
                                               i
                                                     i,L
                                                          i,U
   222   223   224   225   226   227   228   229   230   231   232