Remote Monitoring of Surface Roughness on Turned and Milled Carbon Steels with High-Speed Steel and Tungsten Carbide Tools

Abstract

Machinists and Metrologists are most often called to take decisions for least machining to achieve specific surface qualities of mechanical organs used in aeronautics and other severe environments. The highest challenge they face is in decision-making about the surface finish related to the cutting conditions, the tool life, and the machined material. This paper proposes a method of remote measurement and prediction of the surface roughness on turned and milled carbon steels with high-speed steel and tungsten carbide tools, based on the image acquisition protocol, material content and tool life. The remote measurement of the surface roughness involved a point-to-point viewing angle to capture the image surfaces to appreciate the ideal angle of optimal optical measurement. The assessment of the optical roughness involved the line profiling calculation method on the locally corrected pixels’ values before the areal integration. The optical roughness values were regressed on the reference values and the precision of the method was assessed. The angles of 60°, 75°, and 120° show the effectiveness of the measurement method with precision attaining 83%. For the roughness prediction in the milling and turning operations with high-speed steel and tungsten carbide tools, the fuzzy logic and artificial neural networks techniques are compared considering the cutting conditions as fixed, the carbon percentage and the tool life, all as inputs. With an overall measurement precision above 90% and very low mean square errors, the qualifiedness of the predictive methods is underlined.