Noise-robust tool condition monitoring in micro-milling with hidden Markov models

Abstract

Tool condition monitoring is crucial to the efficient operation of machining process where the cutting tool is subject to continuous wear. In particular, in micro machining, the tolerances, depth of cut, and even workpiece sizes are in micro scale. Micromachining can overcome the shortcomings of micro fabrication techniques (such as lithography and etching) with limitation of work materials (mostly on silicon) and geometric forms (2 or 2.5 dimensions) (Byrne et al. 2003; Liu et al. 2004). One very versatile micro-machining process is micro-milling. Micro-milling has advantages over other micro-machining techniques with respect to the types of workable materials and the free-form 3D micro structures with high aspect ratios and high geometric complexity. However, in micro-milling, with the miniaturisation of the cutting tool (<1 mm in diameter), and the use of high speed (>10,000 rpm), the tool wears quickly. It is critical to monitor the tool wear in micro-machining due to the high precision required. Compared to conventional machining, the noise component in the signal for monitoring micro-machining is usually very high and difficult to separate (Tansel et al 1998; Zhu et al. 2007). This phenomenon makes it difficult to apply TCM in micro-machining. © 2008 Springer-Verlag Berlin Heidelberg.