Research On Fuzzy Diagnosis For Faults In Concrete-Pump

The fault diagnosis process is the requirement of modernization. The concrete-pump becomes an ideal machine of all carrying-concrete machinery, which has been used in wide-ranging, and must be used continually for the construction. The faults of hydraulic system for concrete-pump are random and concealing, which want the repairman having the knowledge and the experience about them, so the skills of repairman become decisive factor. At present as the research of fault diagnosis for it is less, the faults of hydraulic system for concrete-pump are analyzed, neural-network and fuzzy-control theory are introduced into the fault diagnosis process, and used to model the process according to the actual fault phenomenon in this paper. By this means, a feasible approach is put forward for the computer aid fault diagnosis in it. The fault examples of hydraulic system for concrete-pump are analyzed, and their characteristic parameters are chose, so the corresponding relations of the fault diagnosis model are found. Then fuzzy-control theory is used to design the fuzzy-controller, and to model the fuzzy diagnosis process for fault in concrete-pump. Neural-network, which possesses non-linear adaptable information processing ability, and BP NN which has self-learning ability, are used to model fuzzy-neural-network according to the corresponding relations between the chose parameters and the faults in concrete-pump. MATLAB can construct typical NN arousing function. The functions of NN in MATLAB are allocated to model a neural-network according to fuzzy-diagnosis-model, then the neural-network trains by the actual input-output parameters of the model. The accuracy and the adaptability for the model are improved. And the training result of NN could be explained by the fuzzy-diagnosis-model. So the defects in original NN are dispelled whose memory is not easy to understand. In the end, the fuzzy-neural-diagnosis-model is described from the hardware component and the software ability, and emulated by the computer.