Study On Fault Diagnosis Of Two Special Gears

Gear equipment is a key component of the modern industrial transmission equipment. Ithas significant influence on the operation of the whole system. To analyze the workingcondition of the gear system, it is necessary to establish the model of its mechanics and findthe equation of state and the law of motion of the gear system. Gear state detection andfailure prediction can be processed by the theoretical basis above.However, different kinds of gear has different law of motion, the method of analysis ofthe different gear may vary. This article focuses on two commonly used simple gears. One isHardened arc bevel gear, it is often used in complex work environment, the other one iscylinder spur gear which is usually used too. Due to the different physical structure of the twogears, the failure analysis method is also different.Because the tooth surface and tooth line of Hardened arc bevel gear are non-linearequations, the mechanical equation of it is also complicated. This article use artificialintelligence fault diagnosis method to diagnose its failure. Taking into account the hardenedspiral bevel gears have characteristics such as stable operation, long life and low failure ratecompared to the conventional gears, so the support vector machines are used for faultidentification because it is available for small sample fault identification. Because the waveletpacket decomposition can analysis signals by multi-scale and multi-resolution and applies tonon-stationary gear fault vibration signal analysis, this paper adopt the wavelet packetdecomposition energy and their variance as the input of the support vector machine. Throughthe analysis of the experimental results, this method has high recognition rate in smallsamples. It is a viable hardened spiral bevel gear fault diagnosis method.Cylinder straight gear Vice, its physical structure is simple and its fault state can bepredicted by bond graph method. Power bond graph method can describe many complexsystems in vector form by bond graph components. It has been used in the field of mechanical,thermodynamics, physiology, chemistry, fluid, biology, electronics, and magnetic systemsengineering and technical field. The fault tree model established by bond graph can analyzethe gear failure simply, you can visually analyze mechanics, and to be able to push to theequation of state of the gear transmission, using of the computer can quickly establish a bondgraph model and fault tree of a special failure phenomena, which greatly shortens the time ofthe fault modeling. Results of simulation shows that failure phenomena of lower torque of thedriven gear was caused by the failure of the gear friction coefficient decreases, which issimilar to the diagnostic failure results based on fault tree. So it is a viable cylinder straightgear fault diagnosis method.