抄録
It is important for power system operators to determine the optimum network configuration. This paper proposes a new theoretical method to derermine the optimum radial network configuration of a power system.
This algorithm consists of processes as follows, (1) a power system equation can be represented by using node-flow matrix based on Kirch Hoffs 1st law, (2) after diagonalizing the matrix, branches related to non-basic variables of simplex matrix are regarded as open ones, while branches related to basic variables are regarded as close ones of a radial network, (3) branch on-off operation corresponds to the pivot operation of LP method, (4) the objective function is renewed on every pivot operation.
By using this method on a distribution power system with 539 branches, it is found that calculated results are practically useful. And the calculation time (11 seconds) is also practically satisfied by taking measures as follows, (1) pivot selection only from remote controllable branches, (2) per unit description and sign reverse operation of branch flow. Moreover, even if the calculated value is the minimum solution by the effects of blanch capacity constraints, quasi optimum solution is obtained by using double pivot operation. This logic can be applied to not only a distribution network but also a transmission network at any system state such as normal situation, outage, maintenance, and restoration.
