Research On Cutting Tool Wear Condition Monitoring Based On Machined Surface Image

It is very important to improve the automation machining efficiency and quality through the monitoring technology of cutting tool. Regard the machined surface image in turning processing as an object, the internal relation between the tool wear and the micro-images on the machined surfaces is analyzed, the key monitoring technology based on tool wear condition of images on the machined surfaces by the method of image processing, Hough transform, fractal theory and fuzzy pattern recognition in the paper, which provides a new approach for tool condition monitoring.The wear process and condition of the tool are discussed. The mapping relation between the tool's geometry shape and the texture morphology on the machined surfaces is analyzed. The experimental system based on the machined surface images of tool wear monitoring is built, meanwhile, the feasibility, reflecting the wearing condition of cutter through the image of machined surface indirectly, is verified on the basis of analyzing the experimental data.Aiming at machined surfaces images, some technology about image prep-processing is researched, such as image cutting, the correction of non-uniform illumination, image smoothing, edge detection, the correction of image's texture's angle. The algorithm of image processing is done, which lays a foundation for image feature extraction of the monitoring of wearing condition.According to the variation characteristics of the edge images with the tool wear, the Hough transform is applied to analyze and detect the distribution features of line segments in the edge images; the variation laws of average length of line segments and the average angle between line segments and direction of cutting velocity with the tool wear are researched. The experimental results show that the closer correlations exist between the tool wear condition and the two feature parameters; the tool wear condition monitoring can be realized based on the variation law of the feature parameters. In the basis of researching the method of fractal dimension, the concept of the wavelet fractal dimension and the fractal spectrum dimension is put forward through importing the fractal theory into image analyzing on the machined surface, meanwhile ,the particular method about the two dimensions is carried out. The changing law of wavelet fractal dimension and fractal dimension with tool wear condition is researched. The experimental results show that the closer correlations exist between the tool wear condition and the two dimensions; the tool wear condition monitoring can be realized based on the variation law of the two dimensions.Fuzzy pattern recognition is applied to monitor the tool wear condition based on the close connection between every characteristic parameters and the tool wear condition.Building the fuzzy pattern systems of tool wear condition monitoring, every characteristic parameters which has close connection with tool wear condition is input to fuzzy distribution function.Studying the tool wear condition membership of every experimental samples,the tool wear condition identification can be realized based on the maximum subordination principle.The result shows that this fuzzy pattern systems can be used to identify the tool wear condition effectively.