Research of shield machine fault prediction system based on improved elman network algorithm

As influenced by the geological complexity and many uncertain factors, tunnel construction using shield machine is still a complex and risky activity. When passing through the transition geological layer or occurring error operation of the machine, usually some construction failures may occur, such as caking, soil occlusion, water spewing and ground surface subsidence, etc. In order to effectively avoid the failure of construction, this paper proposes using the improved Elman Neural Network Algorithm and making on-line fault diagnosis prediction according to the state parameters of shield machine. Through an engineering instance data normalized simulation study, the feasibility of the system design is verified.