In-Process Detection of Tool Failure in Milling Using Cutting Force Models

Abstract
The authors show that first and second differencing of a time averaged resultant force is extremely effective in the recognition of tool breakage in milling. In order to be useful in production, however, such a system needs to have a knowledge of allowable levels of the first difference. It is shown that these levels, (thresholds), are extremely dependent on the ratio of cutter radius to width of cut, (immersion ratio). A method of on-line identification of the immersion ratio and threshold tuning is presented with experimental verification.