Up to date, the direct nonlinear predictor-corrector primal-dual interior point algorithm (PCPDIPA) is recognized as an effective method for many power system optimization problems. Its efficiency depends on sparsity techniques and a dynamic estimation scheme to decrease the so-call barrier parameter. However, if the value of the barrier parameter is decreased too fast, it would result in small step sizes that halt the convergence of PCPDIPA. In this paper, a hybrid method is proposed to tackle this difficulty. Numerical results of several different sized power systems are presented to illustrate the performance of the proposed method. It is found that the hybrid method reduces the number of iterations by 9%25% and the CPU time by 6%22% for all test cases.
|Number of pages||10|
|Journal||Journal of the Chinese Ceramic Society|
|Publication status||Published - 1996|