An efficient predictor-corrector primal-dual interior point method (IPM) is
applied to reactive power optimization. A linear model is introduced in wh
ich control variables and voltage increments are linked by a modified Jacob
ian matrix, and transmission losses are represented as a function of voltag
e increments. Since this method does not calculate any matrix inversion or
second-order derivatives of Hessian matrix at each iteration for solving th
e linear program, its solution will save computation time and memory space.
Numerical results show that this technique is very effective and fast for
large reactive power optimization. (C) 2000 Elsevier Science Ltd. All right
s reserved.