The purpose of this study is to develop an algorithm to accurately detect faults in series capacitor compensated (SCC) power transmission lines. The line fault must be distinguished from stable power swing, compensating unit malfunction and defects on other lines sharing the same bus (external faults).
In this regard, an effective fault feature extractor based on the cumulative sum (CUSUM) of the amplified second harmonic of the phase currents is suggested. The features are then applied to an artificial neural network for classification. No-fault cases include stable power swing and several disturbances. Due to the independent analysis of each phase, faulty phase detection is also a by-product.
Various fault scenarios are defined, and the algorithm success rate is compared with some newly published methods. Extensive simulations performed over a single-machine infinite bus, a 3-machine, 9-bus and the large-scale New England IEEE 39-Bus networks all indicate that the proposed algorithm can trip the faulty line more quickly and accurately than the contestant algorithms.
Suggestion of a new algorithm based on the CUSUM of the amplified second harmonic of the phase current for the fault feature extraction that is able to isolate the transmission line internal faults from stable poser swing, line compensating unit malfunction and faults on the adjacent lines connected to the same bus.
