Reliability of Monitoring Signals for Estimation of Surface Roughness in Metallic Turned Parts
Current trends in machining processes are focused in three goals: to increase the productivity and the reliability and to minimize costs. In this context, the development of signal monitoring systems is of vital importance for surface roughness inspection. One of the research lines associated to this context is oriented to predict surface roughness using indirect signal analysis, such as cutting forces or vibrations in the machining process. This paper analyzes the results obtained when comparing different nature signals combined with cutting parameters. The final goal is to quantify the deviations obtained with different monitoring signals for establishing the best ones to use as roughness evaluators. The best predictions were obtained when force and cutting conditions were combined together. The absolute error values remains always below 1.28 and 1.11 µm when using the median and root mean square (RMS) as descriptors, respectively.