Page 212 - Mathematical Models and Algorithms for Power System Optimization
P. 212

Discrete Optimization for Reactive Power Planning 203

                       Table 6.20 Basic calculation conditions for the algorithm based on expert rules
                                            The Number of
                The Number of  The Number of  Variable Tap
                Nodes            Branches      Ratios            The Number of Capacitor Nodes

                230                270          105                         59
                                                          Original nodes  Original+new  New nodes
                                                               30           10            19



                                    Table 6.21 Comparison of constraint violations

                                    Expert System2Violation Value  Only Rounded-Off2Violation Value
                Case Name           ΣΔU (p.u.)     maxΔU (p.u.)      ΣΔU (p.u.)     maxΔU (p.u.)
                951D                  0.2409          0.0257          0.2954           0.0207
                951X                  0.2108          0.0172          0.3148           0.0310
                952D                  0.0464          0.0097          0.3731           0.0279
                952X                  0.2736          0.0221          0.2826           0.0236



               Fig. 6.10 shows the comparison between the original and the new capacitor bank number.
               To simplify, the system nodes are ranked in such a sequence: nodes 1–31 are existing
               capacitor nodes, nodes 32–43 are mixed nodes with both old and new capacitors, and nodes
               44–64 are new capacitor nodes. According to the optimization calculation results, more
               existing capacitors shall be installed in nodes 1–31, less new capacitors installed in nodes
               32–43, and fewer new capacitor nodes installed in nodes 44–64.
               The partial planning results of large mode and small mode in winter-summer in 1995 are

               listed in Tables 6.24 and 6.25, respectively. In the tables, ΔP z is total grid loss (MW), ΔP 110
               is 110kV grid loss (MW), and COST is investment (10,000 yuan).

               Tables 6.24 and 6.25 show that the results of VAR optimization procedure are much better
               than manual planning results. With investment minimization as the objective, investment
               costs are greatly reduced. Moreover, grid loss is also less than that with manual planning.
               If power loss minimization is taken as the objective, grid loss will be further reduced. However,
               the cost of investment on VAR units will be increased. The objective of investment
               minimization can help to reduce initial investment but lead to a higher operating cost in a
               long-term operation. By contrast, with power loss minimization as the objective, operating
               conditions will be improved, and long-term operating costs will be reduced. However,
               initial investment has to be increased. How to make a trade-off between the two factors?
               Planning personnel can perform further economic and technical analysis according to
               capital source, operating characteristics, and manpower allocation, etc.
   207   208   209   210   211   212   213   214   215   216   217