A　method　for　fault　diagnosis　of　three-phasesasynchronism　switching　for　a　low　voltage　circuit　breaker　(LVCB)　is　concerned.　Firstly,　the　vibration　signal　is　decomposed　into　several　intrinsic　mode　functions　(IMF)　by　empirical　mode　decomposition　(EMD).　By　analyzing　the　vibration　signal　spectrum　of　a　LVCB,　the　front　four　IMF　components　were　determined　as　the　vibration　signal　characteristic　so　that　noise　of　vibration　signal　was　eliminated.　Secondly,　the　correlation　dimension　of　front　four　IMF　components　was　calculated　by　fractal　theory,　that　is　the　fault　characteristic　of　three-phases　switching　asynchronism　of　a　LVCB.　Finally,　extreme　learning　machine　was　introduced　to　build　the　fault　identification　model　of　three-phases　switching　asynchronism.　Results　of　experiment　and　simulation　showing　that　it　is　effective　to　identify　switching　synchronism　with　ELM　model　based　on　EMD　and　fractal　theory.　In　addition,　the　method　also　has　the　feasibility　to　diagnose　other　faults　of　a　LVCB,　based　on　above　fault　diagnosis　principle.　©　2016,　Harbin　University　of　Science　and　Technology　Publication.　All　right　reserved.