Initialisation, decoupling, hot start, and warm start in direct nonlinear interior point algorithm for optimal power flows

Effective approaches dealing with four important issues in a direct nonlinear interior point algorithm (DNIPA) for optimal power flows (OPF) are presented. The first issue, initialisation, is important for the robustness of DNIPA, the second issue, decoupling, is important for OPF applications. The other two issues are particularly important to OPF performance in energy management systems. The approaches enhance the performance of DNIPA for OPF problems in terms of iteration counts and CPU times. Extensive numerical tests are performed on two different sizes of real-life systems: 244- and 2423-bus systems. Based on these results, the initialisation approach improves direct nonlinear predictor-corrector primal-dual interior point algorithm (PCPDIPA) by 24%, in term of CPU times. The decoupling approach enhances PCPDIPA (with initialisation) by 27%,. The hot start and warm start solve the problems about 2.0 and 2.7 times, respectively, faster than PCPDIPA with initialisation, and 2.5 and 3.3 times, respectively, faster than PCPDTPA without using initialisation