Extracción de rasgos de las señales para la monitorización indirecta de la herramienta en el microtaladrado

Abstract

The paper presents a study on the signals of a microdrilling process in order to extract features that can be correlated with the cutting tool condition, providing the foundations for further developments of indirect cutting tool monitoring systems. In a microdrilling process of a tungsten-cooper alloy, with TiAlN-coated tools, signals of forces and vibrations were measured online. Three different tools with diameters of 0.1, 0.5 and 1.0 mm, respectively, were used and five consecutive holes were elaborated with each tool. In each hole were measured not only the above- mentioned signals but also the temperature and dimensional variations of the cutting tool. Measured signals were processed by using time-domain statistics, Fourier fast transform and Hilbert-Huang transform for extracting features. Correlation analyses between the obtained features and the number of elaborated holes were then carried out, in order to identify which of these features can be used for estimating the tool condition. Time-domain statistics features did not show remarkable correlation with the tool usage level. On the contrary, fast Fourier transform and Hilbert-Huang transform yielded very interesting outcomes, because some of the analyzed features showed a clear elationship with the number of elaborated holes. Future research will be focus on the integration of the different signal processing and decision-making strategies by using artificial intelligence techniques such as neural networks or neuro-fuzzy systems.