Statistics and Its Interface

Volume 1 (2008)

Number 1

Spectral change detection for deep-hole drilling

Pages: 115 – 123

DOI: https://dx.doi.org/10.4310/SII.2008.v1.n1.a10

Authors

Anita S. Busch (Corporate Quality Management & Optimization Methods, Hella Corporate Center, Lippstadt, Germany)

Ursula Gather (Fakultaet Statistik, Technische Universitaet Dortmund, Germany)

Abstract

A control chart for detecting changes in the spectrum of a time series is developed. It is based on a spectral representation using the SLEX function. The control chart compares spectral characteristics from recent to past time windows with spectral divergence measures. As an application, the structure-borne sound measured at the single-lip deep-hole boring machine is put under statistical investigation in order to detect signs of chatter vibration in the process. Different spectral divergence measures are empirically compared with respect to their ability to separate non-chattering from chattering process parts.

Keywords

chatter detection, control chart, exponential weighting, smooth localized exponential, spectral divergence measures

Published 1 January 2008