Vibration Analysis And Fault Diagnosis For Fault Gear

As an extremely wide used component of coupling and transmission in modernmechanical equipment, the gear has a high fault probability. Its fault will cause an outage,shutdown, and even a threat to the safety of personnel. There are a lot of gear faultdiagnosis methods, but the traditional methods don’t take the nonlinear factors intoaccount when making the fault diagnosis. In order to compensating for the limitations inthis regard, based on the research achievement of previous study, this paper presents a newgear fault diagnosis method which combines the nonlinear branch discipline---the fractaltheory and neural network which has nonlinear characteristics.Firstly, this paper introduces the research significance and development of the gearfault diagnosis, analyzes the deficiencies of the traditional methods of fault diagnosis, andthen presents the research content and method.For the gear fault, this paper does research and introduces of the classical faults,analyzes the types of gear vibration when running, and chooses three classic types of gearfaults to analyze its vibration characteristics by the finite element method. Then this papersimulates classic faults of gear on homemade test bed to acquire the vibration signals ofthe gear.Based on the fractal theory, the vibration signals are processed by the fractal filteringand the fractal dimensions calculating. Pattern recognition is done via BP neural networkwhich its inputs are the fractal dimensions and statistical parameters of the vibration signal.This realizes the combination diagnosis via the fractal theory and neural network. By thedouble verification of theoretical data and experimental result, and the high diagnosisaccuracy rate, this method proves to be correct and validated.A gear fault diagnosis system is designed by the Visual Basic and MATALB, whichhas a good user interface and easy operation. This system makes a summary of theresearch content, which makes this paper advance from theoretical research towardspractical application.