A Fresh Way Targeted on Gear Malfunction Detection

As the most common components in the transmission system of machine, gears play an extremely vital part in it. However, it is precisely because of the key of the gear that the failure of the gear in the mechanical system has become an vital part that cannot be ignored. Gear failures may occur due to production quality, transmission maintenance, or inevitable use and wear. When a gear failure occurs, the specific cause of the failure is unpredictable and occurs randomly. In actual engineering, it is not easy for us to check and determine the failure type of the gear in time. Therefore, an optimization way is urgently needed to timely detect and judge the malfunctions of gears in actual production. It can be known from practical experience that the malfunction of the gear will cause the system to show a unique vibration signal, which can be used to judge the malfunction of the gear. Thus, this paper simulates the occurrence of malfunction signals on the basis of finite element modeling to generate test samples for extreme learning machine training. The experimental results show that the fresh gear malfunction detection way is feasible and achieves the expected accuracy.