The invention discloses an
electric locomotive idling identification method based on empirical
wavelet Hilbert transform. The method comprises steps of performing empirical
wavelet transform and Hilbert transform on the locomotive wheelset speed through an idling trend
recognition system to obtain a Hilbert
energy spectrum, and further obtaining a current idling trend value of the locomotive through idling trend recognition based on a time frequency-
energy spectrum; meanwhile, obtaining a current working condition
information value through a traction moment instruction; and finally, comprehensively judging the idling state of the locomotive by combining the current idling trend value, the current working condition
information value and a differential idling judgment value judged by a
differential threshold method, and identifying the idling state of the locomotive. Compared with the prior art, the method has the advantages that the time-frequency information of the input
signal is comprehensively utilized, and idling recognition is more accurate and faster; the adaptability to complex operating conditions and operating environments of electric locomotives is higher; the value requirements on threshold values such as
gain coefficients are not strict; interference of
noise signals can be effectively filtered out, and key signals containing idling are extracted; interference of
train vibration on a
recognition algorithm is effectively avoided, and idling recognition precision is improved.