Study on the Neural Network Model for Shield Construction Faults Diagnosis

In order to solve the problem of establishing the mathematic model for shield construction faults diagnosis, an approach to the mathematic model by using BP neural network is presented in this paper. The BP neural network model for diagnosing three familiar shield construction faults based on the data of shield excavation parameters was built. The inputs of the model are respectively nine shield excavation parameters which are correlative with shield construction faults. The outputs of the model are three shield construction faults which are respectively the spewing at screw conveyer, the wear of disc-cutters and the jamming of shield. The case study of a shield project validated that the structure of the established model is practical, the diagnostic results are right and the diagnosis method is effective. The conclusion provides the beneficial guidance for the design of the online diagnosis system of shield construction faults based on the data of shield excavation parameters.