Terminal Fault Diagnosis For Hydraulic System Vehicle

Hydraulic system as the core component of construction machinery, itsstructure is complex, fault location concealment, after the event of failure willaffect the normal operation of the entire machine. With the traditional faultdiagnosis technology is difficult to accurately identify the fault. How the faultdiagnosis system running real-time monitoring and fault diagnosis fast is thecurrent hot research.Based on study of traditional fault diagnosis technology and embeddedtechnology, the proposed design of embedded fault diagnosis terminal,intelligent fault diagnosis technology and embedded technologies, constituteintelligent fault diagnosis terminal. The terminal has small volume, highperformance, low power consumption, portable and so on, and can be easilyinstalled on the engineering machinery.The main contents are as follows:1.In this paper, the development of hydraulic system fault diagnosistechnology and embedded technology are introduced, and in detail elaboratedthe Tiny6410hardware platform, the improvement of BP neural network, andthe3G network module hardware design, the interface of fault diagnosisdesign,.2.In the process of building a fault diagnosis system with BP neural network,we found that BP network using gradient descent method to optimize thenetwork training. But the defect of this algorithm is easy to fall into the presenceof local minima in the network learning and training process. Therefore, wepropose the use of PSO with a strong global optimization ability on BP networkto optimize the weights and thresholds. And aiming at the PSO algorithm existsof premature convergence, low convergence precision and slow convergencespeed, this paper proposes a adaptive chaotic PSO algorithm, so that theaccuracy and speed of convergence of PSO algorithms have been effectivelyimproved, and effectively avoid the premature convergence.Then using theimproved PSO algorithm to optimize BP neural network, so an improvedPSO-BP neural network diagnostic algorithm has proposed.3.Later,we construct embedded fault diagnosis vehicle terminal based on BP neural network, the construction of a BP fault diagnosis system and thetransplant on the ARM,and design well graphical user interface.The real-timemonitoring and fault diagnosis terminal can be achieved operational status of thehydraulic system, and in order to improve the fault diagnosis system ability,youcan update the parameter data of BP diagnosis network through the3G networkor serial.