As the universal parts of electric locomotive connection and power transmission

gear box is the key to the safe operation of electric locomotive.Working principle of RDEAES and ferrography are described.The RDEAES measured a variety of trace elements concentration in gearbox oil

the ferrography measured Dl and Ds in gearbox oil.According to the RDEAES and Ferrography data research

the common fault of the gearbox is determined.Based on the superiority of BP neural network in the fault diagnosis

BP neural network structure and learning rules are introduced.Various types of faults data from Xining locomotive depot are utilized to build a diagnosis of modified BP neural network model

the simulation analysis for test data is realized.Results indicate that the means can be effectively to diagnose the types of Electric Locomotive gear box.