DOI 10.1007/s00170-004-2376-0 ORIGINAL ARTICLE Int J Adv Manuf Technol (2006) 28: 456–462 A.K. Singh · S.S. Panda · S.K. Pal · D. Chakraborty Predicting drill wear using an artiﬁcial neural network Received: 15 April 2004 / Accepted: 12 August 2004 / Published online: 25 May 2005  Springer-Verlag London Limited 2005 Abstract The present work deals with drill wear monitoring using an artiﬁcial neural network. A back propagation neural net- work (BPNN) has been used to predict the ﬂank wear of high- speed steel (HSS) drill bits for drilling holes on copper work- piece. Experiments have been carried out over a wide range of cutting conditions and the effect of various process parameter like feedrate, spindle speed, and drill diameter on thrust force and torque has been studied. The data thus obtained from the experi- ments have been used to train a BPNN for wear prediction. The performance of the trained neural network has been tested with the experimental data, and has been found to be satisfactory. Keywords Artiﬁcial neural network · Drilling · Flank wear 1 Introduction Drilling is one of the important machining operation extensively used in manufacturing industries. Tool wear has a signiﬁcant in- ﬂuence on the performance of a machining operation. During machining, tool wear affects the tool life and surface ﬁnish of the machine component. In the case of drilling, wear is categorized as ﬂank wear, chisel wear, corner wear, and crater wear. Wear on the drill has a deﬁnitive inﬂuence on hole quality and tool life of drill bits. Therefore, online monitoring of drill wear is a very im- portant issue in manufacturing industries, and thus an emergent area of research. Many works have been reported in the broad ﬁeld of tool condition monitoring. The literature is rich with relevant studies: Noori-Khajavi and Komanduri [1] developed a model for online tool wear moni- toring of a drilling operation and observed that only one signal A.K. Singh · S.S. Panda · D. Chakraborty (✉) Department of Mechanical Engineering, Indian Institute of Technology, Guwahati, 781039 Assam, India E-mail: chakra@iitg.ernet.in S.K. Pal Department of Mechanical Engineering, Indian Institute of Technology, Kharagpur, 721302, INB, India is sufﬁcient to monitor the tool wear. Lin and Ting [2] used the force signal to monitor online drill wear. They used the least square method for determining the thrust force, and torque as a function of spindle speed, feedrate, drill diameter and average ﬂank wear. In another work, Lin and Ting [3] used a back propa- gation neural network (BPNN) with sample and batch mode, and observed a faster convergence of error in the case of the sample mode. They also observed that for a neural network with two hid- den layers with same number of nodes, convergence is achieved faster than that with one hidden layer, and they reported that at higher learning rate, error is reduced. Also, Das et al. [4] used a back propagation algorithm for measuring the ﬂank wear of carbide tool in turning operation. Lee et al. [5] used abductive network modeling for drilling pro- cess for predicting the tool life, tool wear and surface rough- ness. The network has number of polynomial functional nodes. Optimal network architecture is prepared based on a predicted square error criterion. Choudhury et al. [6] developed a three- layer, feed-forward back propagation neural network for pre- dicting ﬂank wear in turning operations. They used the geomet- rical relationship in correlating the ﬂank wear on cutting tool with changes in workpiece dimensions. Kosmol et al. [7] used the ﬁnite element method in calculating the wear of drill point, roughness and erroneous surface of bore hole. Liu et al. [8] used the algorithm for the synthesis of polynomial networks for pre- dicting (ASPNS) the corner wear in drilling operation. Li and Tso [9] used the regression model for monitoring the tool wear based on current signal of spindle motor and feed motor. Choud- hury and Raju [10] developed a regression model to measure the ﬂank wear and corner wear of a drill bit in a cutting operation. Tsao [11] used the radial basis function network (RBFN) and adaptive-based radial basis function network (ARBFN) to pre- dict the ﬂank wear in both the cases, and compared their result with experimentally obtained values. Other studies have used the evolution strategy and genetic and other types of algorithms in dealing with tool wear. Davim and Antonio [12] used the evolution strategy for identifying the type of wear in polycrystalline diamond (PCD) drill bits with a metal matrix composite as the workpiece. They used the Pareto opti-