STATISTICAL MONITORING AND CONTROL OF TOOL WEAR PROCESSES

Abstract

Statistical control charts have been successfully used in industry for monitoring stable processes. However, processes with uncontrollable but acceptable trend are common in practice. One typical example is the wear process of cutting tools. Conventional control charts may not serve the purpose of process monitoring. In this paper, a forecast-based technique using Double Exponential Smoothing is proposed. It eliminates the trend component, and control charts are applied to the residuals. Furthermore, a procedure based on double control lines is suggested and adopted in tool wear process monitoring to integrate statistical and engineering properties for better decision making on tool wear-out. Other than the monitoring of tool wear process, the method can be used for better monitoring of other processes with trend. An actual tool wear data set is used as illustration.