An Event Set Mode Recognition Method for Fault Diagnosis of Inverter Output Overcurrent in Electric Traction System

The inverter output overcurrent (IOO) is a common abnormal state in an electric traction system (ETS), which can be caused by several sources within the ETS and results in an immediate inverter shutdown and requires maintenance at the station. In order to improve the efficiency and intelligence level of maintenance, this article proposes an event set pattern identification method to trace the causes of IOO faults. First, according to traction motor control theory and overcurrent fault mechanism analysis, the several feature indicators of related system signals within the period of stator current are extracted by using adaptive sliding window technique. Second, based on historical data, an event generation method is investigated with hysteresis comparators by studying the variation laws of feature indicators, and event template sequence patterns are constructed to represent different fault causes. Then, the online diagnosis framework has been proposed, which includes three parts, namely, diagnostic enabling, event sampling sequence calculation and fault identification. Finally, field experiments are conducted, and the results show that the proposed method has excellent performance under changes in stator current frequency and operating conditions, and it also has a potential for practical applications.