Tool Condition Monitoring and Diagnostics

A smart machine tool must be able to monitor its condition and report problems. Every CNC machine should provide an alarm for tool condition problems, including tool wear, much as every automobile has a low oil indicator. Tool wear is particularly important for unattended machining as a worn tool can ruin a part. All current tool condition systems operate "blind" without direct information on current cutting conditions. Lacking such information, they are not reliable. Normal cutting conditions can cause sensor output to exceed a pre-set threshold, resulting in false alarms. This tool monitoring system will, for the first time, integrate an NC verification system with sensor data to provide a set of "eyes" distinguishing normal changes in sensor output from abnormal events warranting alarm conditions. We will investigate spindle current as an inexpensive and robust indicator of spindle torque and tool forces to compare with nominal (sharp) tool forces predicted by an enhanced NC verification program. That comparison provides a cutting condition-independent discriminator for tool condition monitoring and diagnosis. Extensive experimental validation will test the reliability of spindle current as a force indicator and the comparison of predicted sharp tool forces versus measured forces as a measure of tool wear.