Gear failures are commonly caused by wear and fatigue and past research on gear failures has primarily focused on individual failures. A gear testing machine was developed to explore the mechanism of gear wear-fatigue interaction, and four sets of wear-fatigue interaction tests were performed by changing the operating conditions of the gear drive by regulating the quantity of lubricant in the test gearbox. The K-S test method was used to investigate the gear life distribution under wear-fatigue interaction. The frequency domain properties of the wear-fatigue interaction were compiled and analyzed using Fast Fourier Transform (FFT) analysis. A coordinate measuring machine was implemented to obtain the exact amount of wear at each stage of the gear test. The mechanism of gear wear-fatigue interaction was analyzed by scanning electron microscope observation of tooth surfaces. The results are as follows. The life distribution under wear-fatigue interaction obeys a log-normal distribution. With the increase of the lubrication the pitting fault occurs, the fault gear rotation frequency and its harmonic frequency appear in the low frequency region, and its amplitude increases with the increase of the pitting fault. In the high-frequency region, a sideband with the faulty gear rotation frequency as the interval appears around the meshing frequency and its two times frequency, and its amplitude increases with the increase of the pitting fault. With the decrease of lubrication, the gear wear failure increases, the amplitude of the meshing frequency and its two times frequency increases in the high frequency region. In the gear transmission process, wear and fatigue exist in a competitive relationship of mutual constraints.

The calculation of tooth wear under mixed elastohydrodynamic lubrication is very complex and requires consideration of many conditions such as load distribution in the tooth meshing zone, micro-convex elastoplastic deformation and tooth surface temperature. The accurate calculation of tooth wear requires a lot of time and effort. In order to calculate tooth face wear under mixed elastomeric flow lubrication quickly and accurately, a new proxy model of tooth face wear is developed using the Kriging method. The pressure distribution required for the wear calculation was obtained utilizing the modified Reynolds equation and ZMC elasto-plastic model. The numerical calculation model of gear wear was derived using the modified Archard wear model. The Kriging model was used to construct a proxy model between gear parameters and tooth wear, and the degree of approximation and goodness of fit of the Kriging model were investigated. The results are as follows. The wear depth at each position is different, the smallest at the pitch, the largest near the tooth root, and the pinion has a larger wear depth than the gear. The Kriging model is highly efficient and accurate in its computation and overcomes the shortage of excessive time spent on the calculation of numerical calculation models.