A new strategy for tool condition monitoring of small diameter twist drills in deep-hole drilling Robert Heinemann n , Sri Hinduja School of Mechanical, Aerospace and Civil Engineering, The University of Manchester, Sackville Street, Manchester M13 9PL, UK article info Article history: Received 21 June 2011 Accepted 6 September 2011 Available online 22 September 2011 Keywords: Small deep holes Twist drills Tool life Tool condition monitoring abstract Tool condition monitoring (TCM) systems employed in industry are mostly used to detect tool fracture, although the prevention of it should be the principal aim. This would not only allow for the avoidance of any fracture-related damage to both the workpiece and machine tool, but also for recondition the tool for further use. This paper presents a strategy, which utilises several features extracted from the spindle power and acoustic emission (AE-RMS) signals recorded when drilling small deep holes using twist drills in order to predict an imminent tool failure. A key to achieving this is the subdivision of the drilling cycle into sections and only monitoring those sections in which the most significant change occurs over the tool life. By doing this it is possible to identify the final (i.e. tertiary) tool life stage and replace the worn out tool shortly before fracture occurs, thus improving the overall tool utilisation to. Of 24 drills tested, the TCM system was able to utilise an average of 84% of the tool life; in only one case it failed to detect tool breakage. & 2011 Elsevier Ltd. All rights reserved. 1. Introduction The growing use of automatically operating machine tools has necessitated the development of strategies to avoid a machine tool continuing to run after a tool has broken or been severely damaged during the operation, in order to prevent further damage to the workpiece and machine tool. Two different approaches have been used so far: The first one relies on fixed tool change intervals, based on tool life assumptions usually provided by the tool manufacturers or shop-floor tool life tests. In drilling, the success of this strategy appears to be questionable due to strong fluctuations in tool life [1–3], which can be attributed to workpiece material in-homogeneities, irregularities in the cutting fluid supply, chip motion and tool geometry [4–6]. To avoid tool fractures within the selected tool change interval, rather short tool change intervals are chosen, which has a negative impact on productivity. In the second approach the cutting process is continuously monitored by analysing process signals that provide information about the cutting tool’s condi- tion, albeit indirectly. In the case of drilling, the most commonly-used signals for monitoring are torque and thrust force [7,8] due to the direct linkage between the operation of a drill and the level of forces it generates [9–13], as well as acoustic emission (AE). K ¨ onig et al. [2] observed a steep increase in the AE signal at the end of tool life of small diameter drills, which made them believe that this signal is very suitable for TCM. Heinemann et al. [14] pointed out that both spindle power and AE, unlike the thrust force, showed a signifi- cant change with progressing tool life of small diameter drills, in particular the slope of the signal curves. In deep-hole drilling, Kavaratzis [9] and Heinemann et al. [14] observed fluctuations in the process signals, implying this pro- cess’s stochastic nature, where aspects such as chip evacuation have a significant effect. This made Ravindra et al. [15] come to the conclusion that monitoring the AE signal for tool wear detection is less practical, because it is polluted with noise, which affects the quality of the signal and makes it difficult to set up proper correlations between the signal and tool wear. Moreover, when drilling deep holes using twist drills, Dong et al. [16] noticed that the magnitude of the AE signal increased strongly with the borehole depth. This might have been caused by the chips inside the flutes, rubbing against tool and workpiece [17]. The research conducted so far implies that monitoring the tool condition when drilling deep holes is not at all straightforward. In this paper a new strategy is introduced, which aims to identify the final stage in the life of small twist drills when drilling deep holes. Although this new strategy uses process signals and features utilised before, the novelty is that the features are only extracted from certain parts of the drilling cycle. It must be emphasised that this strategy does not aim to capture the Contents lists available at SciVerse ScienceDirect journal homepage: www.elsevier.com/locate/ijmactool International Journal of Machine Tools & Manufacture 0890-6955/$ - see front matter & 2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.ijmachtools.2011.09.002 n Corresponding author. Tel.: þ44 161 306 3809; fax: þ44 161 306 3755. E-mail address: robert.heinemann@manchester.ac.uk (R. Heinemann). International Journal of Machine Tools & Manufacture 52 (2012) 69–76