Monitoring and control systems for machine tools are essential for increasing productivity. A robust monitoring system, coupled with the ability to use machining state signals predicted by the digital model is key to the implementation of such systems in production environment. This thesis presents the use of machining simulations to CNC-inherent or accessible data collected from sound, vibration, and force sensors. Through the combination of simulations and on-line measurements, a digital twin is created to detect chatter, tool breakage, and tool wear.

First, the machining process states such as force, torque, power, and cumulative chip removal are simulated along the tool path. The actual and virtual positions of the tool along the tool path are synchronized during actual machining so that measured and simulated states can be compared.

A new tool wear monitoring algorithm is proposed. The cutter – workpiece engagement area and cumulative chip removed by the cutting edge are computed using the Virtual Machining Software developed in the laboratory. The spindle servo motor current is collected from the CNC and normalized by the engagement area. The tool wear is correlated to cumulative chip thickness and an increase in the geometry-independent spindle motor current using a few tool wear measurements. It is shown that the tool wear progress can effectively be monitored by integrating simulation and motor current extracted from the CNC system.

Similarly, chatter is also detected from sound spectrum measurement along the tool path by differentiating it from the air cut, transient vibrations and changes in the workpiece geometry with the aid of digital simulations. Chatter detection and avoidance algorithm is also enhanced by deactivating it at transient cutting zones.

In some applications such as adaptive force control, it is necessary to measure cutting forces during machining. A commercial tool holder equipped with accelerometers is used to predict cutting forces from vibration data. The transfer function between the vibrations measured by the instrumented tool holder and the applied force is modeled. The cutting forces are predicted from the vibration measurements with the aid of Kalman filter and compared against the digital estimations along the tool path.