Real-time failure forecasting for flat, ball-nose, roughing, and tapered endmills using an accelerometer

In this paper, flat, ball-nose, roughing, and tapered endmilling tools are investigated for their respective wear trends. These curves are contrasted with the trend exhibited by traditional flat endmills. To isolate the wear behavior of the various tool geometries, an autoregressive based monitoring algorithm is used to track the tool's condition using a tri-axial accelerometer. An accelerometer is used due to its low cost and since it does not limit the machining envelope. To demonstrate repeatability, eight life tests were conducted (two life tests for each type of endmill). The technique discussed herein successfully identifies impending fracture or meltdown due to wear in all cases, providing sufficient time to remove the tools before failure is realized. Furthermore, the algorithm produces similar trends capable of forecasting failure, regardless of tool geometry, in all cases, without requiring algorithm modifications or prior information regarding the cutting conditions.