International Journal of Machine Tools & Manufacture 44 (2004) 1261–1269 www.elsevier.com/locate/ijmactool Tool life determination based on the measurement of wear and tool force ratio variation S.E. Oraby a,1 , D.R. Hayhurst b, a Department of Production Engineering, Faculty of Engineering and Technology, Suez-Canal University, Port Said, Egypt b Department of Mechanical, Aerospace and Manufacturing Engineering, UMIST, P.O. Box 88, Sackville Street, Manchester M60 1QD, UK Received 19 April 2004; accepted 29 April 2004 Abstract Non-linear regression analysis techniques are used to establish models for wear and tool life determination in terms of the vari- ation of a ratio of force components acting at the tool tip. The ratio of the thrust component of force to the power, or vertical, force component has been used to develop models for (i) its initial value as a function of feed, (ii) wear, and (iii) tool lifetimes. Predictions of the latter model have been compared with the results of experiments, and with predictions of an extended Taylor model. In all cases, good predictive capability of the model has been demonstrated. It is argued that the models are suitable for use in adaptive control strategies for centre lathe turning. # 2004 Elsevier Ltd. All rights reserved. Keywords: Tool life; Tool wear; Tool force ratio 1. Introduction In the manufacture of metallic components using conventional machine tools, cutter tool wear, its effects on surface finish, and cutter tool failure provide major limitations to the achievement of economic production [1,2]. Techniques for describing tool wear and failure are based either on the use of databases with inter- polative and extrapolative techniques, or on the use of explicit mathematical relationships. In both cases, the stochastic nature of tool wear provides an obstacle to the achievement of optimal production conditions. This situation can be improved by using adaptive control (AC) techniques [3,4], which create the possibility to measure and to control the process in situ, and so avoid the need to relate the current machine, workpiece and tool conditions to those conditions under which the existing databases and mathematical equations have been generated. Instead of using complex tool databases, it is pro- posed that models be employed which describe wear; which determine tool life in terms of measurements of in-process variables; and which can, if necessary, be recalibrated during metal cutting. Ideally, the form of the models should be such that the workpiece–tool var- iations, and the variations from machine tool to machine tool, are sufficiently small not to provide obstacles to their effective use in industry. Some poss- ible approaches are now discussed. It is possible to use a range of process variables to indirectly measure the effects of tool wear. Measure- ments have been made by many investigators [5–8]. Theses included the use of on-line electronic equip- ment/transducers to measure: motor power, spindle torque or even the current drawn by a.c. feed drive ser- vomotors [4,9,10]. Another technique which has gained wide acceptance in sensing wear is the dynamic behav- iour of the tooling system [1,11–15]. However, one of the most promising techniques for sensing tool wear and breakage involves the measurement and obser- vation of the components of static/dynamic force act- ing at the tool tip during machining [5,16–23]. In a previous study reported by the authors [24] for centre lathe turning, force variation during cutting was found  Corresponding author. Tel.: +44-161-200-3817; fax: +44-161-200- 4166. E-mail address: d.r.hayhurst@umist.ac.uk (D.R. Hayhurst). 1 On leave. Present address: Department Mechanical Production Technology, College of Technological Studies, PAAET, P.O. Box 42325, Shuwaikh 70654, Kuwait. 0890-6955/$ - see front matter # 2004 Elsevier Ltd. All rights reserved. doi:10.1016/j.ijmachtools.2004.04.018