Fault Diagnosis of Railway Track Circuits Using Fuzzy Neural Network
Ze Liu · Journal of the China Railway Society · 2012
The FNN algorithm was introducd into fault diagnosis of railway track circuits.On the basis of analyzing fault causes and fault features and conforming to the working principle of track circuits,the FNN model was established.The model selected the voltages of three key components as the input and four typical faults as the output.Nine inference rules were summed up by expert experience.The parameters of the FNN model were optimized with the first-order gradient optimization algorithm through learning and training of sample data.Computer simulations show that the FNN model is effective and is of a certain accuracy,and it provides a good method of fault diagnosis of railway track circuits.