Determination of Gear Skiving Tool Life Using an Image-Based Wear Detection System

Abstract

Gear skiving is a high-precision cutting technology that is particularly well-suited for machining internal gears. It is essential to evaluate tool wear on the machine in a timely manner to accurately estimate tool life. Existing research focuses on the development of an evaluation system that estimates wear and breakage through image-based observation. This study outlines the system configuration and presents an image processing method for extracting tool geometry. However, the flank surface reflection interfered with the extraction of the rake face ridge during bottom-up observation, where a reflector was placed behind the tool. To resolve this, a new contour-extraction method was employed. This method involves installing a blue or other colored reflector, illuminating the adjacent surface with diffusely reflected light, capturing an image of the tool bottom illuminated with white light, and applying RGB color decomposition. An index value that comprehensively evaluates the difference between the tool profile before and after each machining operation is also proposed. Additionally, a corresponding procedure is established to identify tool wear or chipping by comparing the index value to a threshold value that is determined based on its relationship to the actual amount of wear. The experimental results demonstrate that the index value increases progressively with tool wear advancement, validating the effectiveness of the proposed method. However, owing to image focus, wear detection can become unreliable, making it a critical consideration in image-based wear detection methodologies.