Statistical quality control through process self-induced vibration spectrum analysis

Abstract

The first part of this study (Carnero et al., Mechanical Systems and Signal Processing, 24:1138–1160, 2010) analysed the influence of the process variables and work cycles on the quality of the bearings manufactured in an automotive bearing plant. The study was focused on the analysis of the overall vibration reading produced by the contact between the tool and the part. An analysis of variance was conducted on the overall vibration readings, which reflected that high-frequency vibration displacements are sensitive to process setup variables as well as the quality of products manufactured. Nevertheless, it was also observed that overall vibration values are not sufficient to analyse the relationship between the mechanical behaviour (vibration) and final quality obtained from high-precision machining processes. In this article, therefore, a new study is conducted based on spectral vibration measurements. A new experiment has been designed taking as input variables the diameter and the rotating speed of the tool. The selection criterion is based on the strong influence of these two variables on high-frequency vibration displacement and quality of parts (chattering, measured in terms of Lob A and Lob B). Two identical grinding machine tools were used during the experimental phase. Output variables are high-frequency displacements and high- and low-frequency chattering. The statistical analysis used in the new experiments determines spectral bands of the process in which vibrations induced by tool–part contact relates to development of lobes in the part to be identified. The study allows identification of vibration bands that are suitable for control in order to guarantee quality of the parts produced. To achieve that goal, the concept of spectral identity of the processes has been introduced to incorporate vibration induced by the process itself in the spectrums and to differentiate that process vibration from other mechanical vibration sources.