The use of Taguchi’s methods for the evaluation of industrial knife design S IMON H SIANG,R AY M C GORRY and ILYA BEZVERKHNY Liberty Mutual Research Center for Safety and Health, 71 Frankland Rd, Hopkinton, MA 0746, USA Keywords: Knife blade; Taguchi’s methods; Knife design. Knives are used in industry, and particularly on meat packaging assembly lines, where a premium is placed on the speed of cutting. One consequence of this is that knife design and selection has been based on the speed of cutting as the most signi®cant measure of a knife. In industry the cuts that knives make vary in diculty with many factors such as thickness, contour, and temperature of the material being cut. These factors may change the speed with which these cuts are made. According to Taguchi’s methods, a robust design is one that maintains high performance while remaining insensitive to changes in its environment. This means that a robust knife design would enable an employee to cut not only with as much speed as possible in one particular cutting scenario, but also with a consistently high speed in as many cutting scenarios as possible. An experiment with various knife designs was evaluated with Taguchi’s method. Ten subjects cut outlined clay patterns in 15 min with each one of nine diŒerent knives. The nine diŒerent designs were generated from the following combinations; three grip sizes (44×4, 50 ×8 and 63 ×5 mm), three coupling angles between the grip and the blade (0, 45 and 90 8 ), three blade heights measured from the middle of the blade (10, 13 and 15 mm), and three blade lengths (85, 125 and 155 mm). Eight conditions of cutting material were used to generate `noise’. They are combinations of temperature of the clay to be cut (from temperature (R) »25 8 C, frozen for 1 h (F) »5 8 C), thickness of the clay to be cut (15 and 25 mm), and orientation of the work surface (0 and 30 8 ). The result shows that the knife design of 63 ×5 mm grip size, 0 8 coupling angle, 13 mm blade height, and 85 mm blade length provides the highest number of cuts, the knife design of 50 ×8 mm grip size, 45 8 coupling angle, 15 mm blade height, and 85 mm blade length gives the highest signal to noise ratio. The shorter blade (85 mm) is suggested by both designs, but the two designs are diŒerent in the other three variables. Further analysis of the noise variables shows that lowering clay temperature signi®cantly reduces cutting speed. Although there is more than one design that can resolve changes in clay thickness and clay orientation, a 15 mm blade height is the only design insensitive to the clay temperature change. 1. Introduction In the competitive meat packing industry knife design is an important factor in yield rate. The high speed of the production chain from which carcasses are hung and the meat trimmed requires a sharp and ecient knife. Furthermore, due to the traditionally high rate of injuries and subsequent employee turnover, knife design plays a substantial role in workplace safety (Cochran and Riley 1986a,b, Hallbeck et al . 1990, OSHA 1990, Riley et al . 1994). A review of data from accidents in the meat packing industry (Pezaro et al . 1985) showed that meat-cutters were injured most ERGONOMICS, 1997, VOL. 40, NO . 4, 476 ± 490 0014±0139/97 $12×00 Ó 1997 Taylor & Francis Ltd