The Empirical Mode Decomposition（EMD） is introduced in this paper to analyze vibration signal of circuit breaker.The energy entropy of Intrinsic Mode Function（IMF） decomposed by EMD is proposed as the new characteristic feature to identify the fault type of circuit breaker.In order to verify the validity of this method

the simulation experiment is carried out on an actual 110 kV circuit breaker in our laboratory.The IMF energy entropy of vibration signal in normal and fault circuit breaker is extracted respectively and is taken as the input of radial basis function neural networks to identify the fault style.The confidence degree is also introduced in this paper to evaluate the output of neural networks.A satisfied result is obtained based on the experimental data

and the proposed method provide a novel way to the fault diagnosis of circuit breaker based on vibration signal.