Page 131 - Modern Control of DC-Based Power Systems
P. 131

Background                                                    95


              error, the larger will be also the compensation term. If y 5 ^y, then the
              state estimation is perfect and there will be no compensation term.
                 In order to determine how quickly the estimated state converges to
              the real state, let us define the error vector as:

                                          e 5 x 2 ^x                     (3.20)
                 By substituting the variables x and ^x from (3.6) and (3.19) and deriva-
              tion, the error vector dynamics can be obtained:
                                       _ e 5 A 2 GCÞe                    (3.21)
                                           ð
                 Therefore, the dynamic of the observer are determined by the eigen-
              values of the term A 2 GCÞ.
                              ð
                 Impact of observer in a closed-loop system with state-feedback: Using the esti-
              mated state of the system (^x) rather than the actual system state (x) in the
              pole placement technique affects the behavior of the overall closed-loop
              system. Using Eqs. (3.4) and (3.19), it can be shown that the dynamics of
              the closed-loop estimated-state feedback system is governed by the fol-
              lowing equation:

                                _ x    A 2 BK     BK       x
                                   5                                     (3.22)
                                _ e       0     A 2 GC     e
                 Consequently, the characteristics equation can be written as:
                                                                         (3.23)
                                              j
                                 j sI 2 A 1 BKj sI 2 A 1 GCj
                 This means that the closed-loop poles of a system with state feedback
              and a state observer are composed of the poles from the state feedback
              when no observer is present plus the poles from the state observer design.
              In other words, the dynamics of the closed-loop system is determined by
              both the poles of the observer and the new poles from pole placement.
                 From a practical point of view, the above conclusion means that the
              observer and the state-feedback can be designed separately. Furthermore,
              in order to minimize the impact of the presence of the state observer on
              the overall dynamics of the systems, its poles are usually placed farther
              from the imaginary axis on the left-hand side plane compared to the
              desired closed-loop poles, which are placed using the state feedback.



                   3.6 DROOP

                   Paralleling a number of converters typically offers advantages over a
              single high power converters, such as low component stresses or increased
   126   127   128   129   130   131   132   133   134   135   136