A machine learning approach to tool wear behavior operational zones

摘要

The range of permitted temperature and stress produced during a machining process is related to the metallurgical properties for each tool material and can be empirically determined. For each combination of tool and workpiece material, particular constants are approximated to prescribe the relationship between the temperature-stress combination and the feed rate-speed combination. Using this concept an operational zone for each tool-workpiece combination can be defined. This paper proposes a machine learning algorithm to determine this operational zone. Instead of relying totally on empirical testing, a learning algorithm is used to incrementally define the operational zone with the related parameters described above. Once determined, the operational zone is then used to enhance machining control.