Harmonic worst-case identification and filter optimal design of MRT systems using genetic algorithms

A mass rapid transit (MRT) operation results in harmonic distortions of AC supply systems. To estimate the effects of harmonics in the system, the worst case should be identified in order to determine whether the harmonic level is beyond the tolerable limit. Levels of voltage and current harmonic distortions in MRT supply systems are affected by train operating modes and system configurations. Harmonic worst-case identification can be treated as an optimisation problem in terms of train separations and traffic conditions. In the paper an approach based on a genetic algorithm (GA) is proposed for solving such problems. The worst-case information is then used in another GA search to place and size the corrective filters when required. Afterwards, the search for the worst case after correction is repeated. The approach uses approximate train movement and consumption models, and the AC/ DC harmonic loadflow for evaluating the overall harmonic distortion. Simulation results demonstrate the robustness of the approach, which is capable of solving highly nonlinear problems beyond the capability of conventional techniques.