Narrow Results

A critical part of a machining system in an unmanned factory is the ability to change the tools automatically due to wear or damage. In order to fit this requirement, on-line real time monitoring of tool wear becomes essential. Thus, this paper is dedicated to introduce some new computer techniques, e.g. artificial neural networks (ANNs), fuzzy logic, and methods, e.g. multi-sensor integration, for monitoring tool wear particularly in turning operations.

In high speed machining, the cutting force changes constantly as the tool becomes worn out. Taking ceramic flank milling as an example, three axial cutting forces were recorded using measuring instruments during the cutting process, with the data forming a time series. Fractal theory was then employed in analyzing these signals. The results show that the cutting force signals of the three axial directions are all kept invariant separately when the cutting process is smooth. However, the situation changes as the cutting tool becomes worn out. Further study indicates that the signal in Y-axial direction changed first as the cutting tool wore out and the fractal dimension of Y-axial cutting signal increases correspondingly. This conclusion could provide a meaningful method for evaluating the cutting quality of ceramics, and can also be useful for assessing the tool wearing in high speed machining.

The monitoring of tool wear state by the texture of the turning surfaces is studied in this paper. A microscopic image acquisition system was established to capture the images from turning surface and tool flank, then the relation between turning surface texture and tool flank wear state was studied based on Fractional Brown motion model. Specifically, logarithm power spectrum (LPS) of each surface texture image was earned by using the Fourier transform, five sets of data was selected from the LPS on five different directions (namely, X-axis direction, 30° direction, 45°direction, 60° direction and Y-axis direction), fit-slope and fractal dimension of earned data were calculated by linear fitting method. Finally, fractal dimension on X-axis direction, found to be highly correlated with the trend of flank wear, can be regarded as the texture feature parameter of tool wear state monitoring.