About this data
Accelerometer (Triaxial ICP Accelerometer by PCB) values are recorded continuously during the process with a 6mm stainless twist drill bit having 4 edges. Sensory data (in all 3 dimensions) is saved into a CSV file. Data has been structured into three classes for flank ware level ("Sharp", "Usable" and "Worn") relative to the process image data collected during the overlapping time period.
To ensure high surface quality and low cost during production milling process, the tool ware has to be monitored constantly. However, manual inspection can be highly time consuming and erroneous. Simple automatic solutions are often not sufficient, since the tool ware behaves usually non-linear and depends on many factors.
The data can be employed in machine learning methods for indirect determination of tool ware / anomalies. The data is generated on a CNC machine in a multi interpolation manufacturing process and 3 tool ware classes were identified and labeled by experts.