Identifying Intermittent Faults to Restrain BIT False Alarm based on EMD and HMM

An intermittent fault identification approach based on empirical mode decomposition(EMD) and hidden Markov model(HMM) to restrain false alarm for built-in test(BIT) caused by intermittent fault is proposed. Firstly, EMD is employed to decompose the original BIT signals which can reflect the fault feature, several energy-dominating intrinsic mode functions (IMF) are selected to extract the feature vector. Then extracted feature value are sent to the trained HMM for state recognition. The results show that the intermittent fault state can be effectively identified by the approach.