Surface roughness prediction based on the correlation between surface roughness and cutting vibrations in dry turning with TiN-coated carbide tools

This paper presents an approach to improving surface roughness prediction that is based on analysing cutting tool vibrations by a novel signal processing technique known as singular spectrum analysis (SSA). Each eigenvalue of the SSA decomposition of each vibration (radial, tangential, and feed vibration) and its corresponding principal component are studied and analysed to select only those eigenvalues of each vibration signal decomposition that contain valuable information for the development of a surface roughness prediction system (SRPS). Also, the influence of tool geometry and tool flank wear on the SSA decomposition of the vibrations is studied. Finally, only the information most correlated with surface quality, extracted by means of SSA of each vibration, is used to develop an SRPS. Experimental results provide conclusive support for the proposed SRPS, and justify the use of the SSA technique in the design of these systems. The ability of the SSA technique to extract only the information correlated with Surface roughness of each cutting vibration and the manner in which tool geometry and tool flank wear influence the final R-a constitute the main contributions of this work.