Abstract: :In order to solve the problem of fault classification on distribution lines， a method was proposed in which atomic energy entropy was combined with continuous string matching (CSM). The fault transient feature was analyzed， the dominant atom and the secondary one of fault transient signal were achieved by atom decomposition method， and the secondary one was used to supplement fault information. Atomic energy entropy of each atom was regarded as the fault characteristic value， representing fault feature more completely. Applying CSM fault classification method could save the time spending on selection and optimization of processing parameters in some smart fault identification method， through the establishment of encoding rules， fault features were encoded， and then the fault type was finally confirmed by searching the encoded fault library. A case study of system model showed that the proposed method is feasible and efficient.

Key words: atomic energy entropy, CSM (continuous string matching), fault extraction, fault classification, distribution lines