Systematic Study Of The Hydraulic Pump Fault Diagnosis Based On Neural Networks

With the development of hydraulic system toward fastness, large power, and high-precision, the functions of hydraulic system and equipments increase, the structure becomes more complex, the performance index requirements are higher and higher and working strength increases, which accordingly increases the possibility of the appearance of hydraulic system failure. In hydraulic system, hydraulic pump is the core of the whole system, also the dynamic source of hydraulic system. It has decisive effect to the whole system. The performance of hydraulic pump may directly affect the life-span of other components and decide if the system can run the correctly. Therefore, for modern industry production, it is significant to research and design the fault diagnosis method of hydraulic pump.This dissertation studies fault diagnosis of hydraulic pump based on neural network. By analyzing fault mechanism of hydraulic pump and the method of fault diagnosis, this dissertation presents a sort of neural network method based on BP algorithmic to diagnose hydraulic pump, and establishes the structure of neural network, learning sample and each parameter for hydraulic fault diagnosis, even trains the network by all of these. The dissertation designs the fault diagnosis system with VC++ 6.0, and implements two modules, neural network learning and neural network diagnosis. This dissertation also presents the adoption of three reasoning mechanisms based on neural network, forward, backward and mixed two-way, which users can choose according to practical demands in the process of diagnosis. The design and implementation of this system makes the fault diagnosis of hydraulic pump effective, and this system is very convenient in practical use.