On-line tool-workpiece contact detection is based on cutting forces signal
Information about tool location according to workpiece is crucial for performing micro-milling correctly. Workpieces prepared for micro-milling processes can have different dimensions after machining operations. There is a need to find the “zero” point of workpiece surface in tool axial direction (Z). Zero point is usually specified as the point of toll and workpiece contact. The easiest (but most time-consuming and most demanding for machine operators) method of finding contact is observation of the rotating tool, which is slowly moved toward the workpiece. Due to very small tool dimensions, this method requires a microscope for tool observation. Automation of this process gives better repeatability and accuracy of tool-workpiece contact detection. Cutting force signals have not been used before for tool-workpiece contact detection.
The main idea of the proposed tool-workpiece contact detection method is based on the how much the cutting force signals increase in Z axis when the tool touches the workpiece. Cutting forces increase for a very short period of time, thus short time signal analysis has to be used. Systems for online tool-workpiece contact detection are based on a diagnostic system previously described [“Real-time diagnostics system for micromilling,” by Bogdan Broel-Plater, Krzysztof Pietrusewicz, and Paweł Waszczuk, Nov. 7, 2012, CE USA]. A block diagram of the system is shown in Figure 1. The system is scalable and can be extended with acoustic emission sensors or acoustic pressure sensors. All analysis must be performed in real time; therefore, National Instruments’ programmable automation controller CompactRio was used for signal processing. Micro-milling machine Aerotech linear drives also have to be controlled in real time; the movement must be stopped immediately after detection of tool-workpiece contact to avoid workpiece damage. Aerotech linear drives can be directly controlled through National Instruments LabVIEW software.
Method and procedure
First, calibration obtains the the coefficient value for the specified workpiece material. The calibration should use a method that detects tool-workpiece contact, such as observation of the rotating tool, which is slowly moved toward the workpiece. Calibration must be done only once for the specified workpiece material.
Spindle rotation must be on during the procedure. The tool is moved toward the workpiece at speed v in step of Δz. During tool movement, the cutting force signal is recorded. Then the root mean square (RMS) value of the cutting force for n signal samples is calculated. The sampling frequency is set to the maximum possible value (51200 samples per second). The algorithm of the procedure is shown in Figure 2. The algorithm was implemented in National Instruments LabVIEW. Crucial for reliable operation of the procedure is setting the right parameters, such as spindle rotational speed, tool speed v, and step value Δz.
The value signal processing method must be resistant to factors such as a high noise level and very low cutting forces. To achieve this at the beginning of the procedure, when there is certainty that the tool is outside the workpiece, the mean value from m root mean square (RMS) reference values is calculated. Then current RMS value from n signal samples is calculated and compared to the reference value calculated outside the workpiece. A comparison is made with the coefficient defined as current RMS value to reference RMS value. When the coefficient value is higher than previously set for the current workpiece, material contact is detected and reference “zero” point is found.
Improved quality, fewer errors
The proposed solution for on-line tool-workpiece contact detection significantly improves the machine operator’s work and eliminates human error. Due to the applicability of the described method and its varieties, three patent applications have been submitted to the Polish Patent Office. The presented solution can be implemented in any CNC machine system, but it is especially designed for micro-milling applications. The issue of tool-workpiece contact detection will be developed in further studies.
– Marcin Matuszak, Msc, is a PhD student at West Pomeranian University of Technology, Szczecin, Poland. His main field of interest is micro-milling processes, especially cutting forces and dynamics. Paweł Waszczuk, Msc, also is a PhD student there. His PhD thesis examines the problem of integrating correcting functionalities for robust control of digital servodrives. Krzysztof Pietrusewicz, DSc, is an assistant professor at West Pomeranian University. All three contribute to Control Engineering Poland. Edited by Mark T. Hoske, content manager, CFE Media, Control Engineering, firstname.lastname@example.org.