Image-based wear state evolution and in-process recognition method for abrasive belt grinding of GH4169

Abstract

Abrasive belt wear is one of the most important factors affecting grinding quality, especially on difficult-to-machine materials such as GH4169, where belt wear is even greater. Currently, there is no unified method for quantitatively characterizing the wear state of abrasive belts. Moreover, due to the absence of a correlation between the wear degree and machining quality, it is challenging to quantitatively categorize the wear stages and offer process - oriented guidance for parameter optimization. Aiming at addressing the aforementioned issues, the image-based methodology for the evolution and in-process recognition method of the wearing status is proposed for abrasive belts grinding. First, based on the industrial robot grinding platform, the abrasive belt wear experiment is carried out, and the evolutionary process of the abrasive belt wear state of GH4169 as well as the influence law of wear on machining quality are obtained. And then, the influence of abrasive belt wear on the material removal depth has been studied. A wear area coefficient-material removal depth model is established from the perspective of microscopic abrasive particles, and the evolutionary law of grinding material removal depth with different wear area coefficients is revealed. A wear state characterization method based on material removal depth is put forward. And then, an image-based on-machine calculation method for abrasive belt wear area coefficient has been established. Finally, an image-based abrasive belt wear state recognition method for the GH4169 robotic abrasive belt grinding is achieved, and the effectiveness of the method is verified through experiments.