Research Journal of Recent Sciences _________________________________________________ ISSN 2277-2502 Vol. 1(10), 19-26, October (2012) Res.J.Recent Sci. International Science Congress Association 19 Effect of Cutting Parameters on Surface Roughness and Cutting Force in Turning Mild SteelRodrigues L.L.R., Kantharaj A.N., Kantharaj B., Freitas W. R. C.2 and Murthy B.R.N.Department of Mechanical and Manufacturing Engineering, Manipal Institute of Technology, Manipal 576 104, Karnataka, INDIA Department of Automobile and Aeronautical Engineering, Manipal Institute of Technology, Manipal 576 104, Karnataka, INDIA Available online at: www.isca.in Received 13th April 2012, revised 23rd April 2012, accepted 3rd May 2012Abstract The purpose of this paper is to study the effect of speed, feed and depth of cut on surface roughness (R) and cutting force (F) in turning mild steel using high speed steel cutting tool. Experiments were conducted on a precision centre lathe and the influence of cutting parameters was studied using analysis of variance (ANOVA)based on adjusted approach. Based on the main effects plots obtained through full factorial design, optimum level for surface roughness and cutting force were chosen from the three levels of cutting parameters considered. Linear regression equation of cutting force has revealed that feed, depth of cut, and the interaction of feed and depth of cut significantly influenced the variance. In case of surface roughness, the influencing factors were found to be feed and the interaction of speed and feed. As turning of mild steel using HSS is one among the major machining operations in manufacturing industry, the revelation made in this research would significantly contribute to the cutting parameters’ optimization.Keywords: Full factorial design, ANOVA, surface roughness, cutting force, interaction effect, adjusted approach. Introduction An area of research interest in turning is the analysis of cutting force, as minimum power consumption is a never ending endeavour. Among the Cutting force, Thrust force and Feed force the former prominently influences power consumption and the most common equation available for the estimation of Cutting force is given by (equation 1): = k × DOC × f (1) Where, DOC = Depth of cut (mm), f = feed (mm/rev), = Specific cutting energy coefficient (N/ mm) According to equation 1, cutting force is influenced by the depth of cut, feed, and specific cutting energy coefficient. A lot of work is in progress to study this influence and construct the models for different tool and work force material so as to optimize the power consumption. Another important parameter of research interest is Surface roughness of the work piece produced, as an optimum surface finish would influence performance of mechanical parts and cost of manufacture2-7. The surface finish of any given part is measured in terms of average heights and depths of peaks and valleys on the surface of the work piece. But there are basically two streams of arguments on the influencing factors of surface roughness. A popularly used model for estimating the surface roughness value is as follows (eqn. 2)9-10. = f/8r (2) Where, R = ideal arithmetic average (AA) surface roughness m), f = feed (mm/rev), r = cutter nose radius (mm).The second stream of argument introduces speed also as an influencing factor of surface roughness and the governing equation is defined as (equation 3)11. a = C x V x f (3) Where, V = speed (rpm), f = feed (mm/rev), C = Constant. However, both the above two streams of arguments explain the surface roughness partially. Hence, there is always a need to go deeper into the investigation of influencing factors of surface roughness, particularly with respect to the interaction effects such as those between speed, feed, and depth of cut for different combinations of tool and work material. Honet et al. reviewed contemporary work on heat generation and heat dissipation in high speed metal turning on coated materials and also briefly reviewed some temperature measurement techniques in metal cutting12. Considerable research effort has been made on the thermal problem in metal cutting, but the accuracy of the readings and the means by which the temperatures are measured are in question13. The unique tribological contact phenomenon, which occurs in metal cutting is highly localized and non-linear, and occurs at high temperatures, high pressures, and high strains. This has made it extremely difficult to predict in a precise manner or even assess the performance of various models developed for modelling the machining process. Even though thermal aspects are equally important, it is beyond the scope of this research. Research Journal of Recent Sciences ______________________________________________________________ ISSN 2277-2502Vol. 1(10), 19-26, October (2012) Res. J. Recent Sci. International Science Congress Association 20 Astakhov and Shvets studied force variations and their effects in metal-deforming technological processes14. They suggest that interaction of the energy waves propagating in the medium might affect the cutting force. They experimented and studied on the interaction between the deformation and the heat waves. The conclusions drawn from this paper reveals that the study of cutting force and the interaction between the deformation and heat waves can be very helpful in adopting the process which involves the least energy consumption. The study on MDN250 steel using coated ceramic tool by Lalwani et al. also portrays the significance of accounting for surface roughness and cutting force15. MDN 250 steel finds its applications in Aerospace industry, Naval industry, etc., and hence, the results and the conclusions drawn here will prove to be helpful in the selection of optimum manufacturing conditions, thus contributing towards larger productivity16,17. Having realized the importance of the choice of most appropriate cutting conditions in metal cutting, this research primarily focuses on machining mild steel using HSS owing to its lower cost, ready availability, and a wide range of applications from automotives to domestic goods to constructional steel and many other machine elements such as keys, rings, fence posts etc. The influence of cutting parameters on cutting force and surface roughness can be studied effectively using Adjusted statistical approach18-21. Material and Methods Machine:The experiment was carried out on the precision centre lathe (PSG A141) which enables high precision machining and production of jobs. The main spindle runs on high precision roller taper bearings and is made from hardened and precision drawn nickel chromium steel. Technical Specifications are: centre height: 177.5mm, main motor power: 3hp, 30 longitudinal and transverse feeds. The Tool:HSS tool with the alloying elements: manganese, chromium, tungsten, cobalt etc. has comparatively better resistance to heat and wear. Tool length of 75mm (approx.) was taken so as to minimize undesirable vibrations, which would influence cutting force and surface roughness. The lathe tool dynamometer was used for measuring cutting force and cutting process was continued until significant tool wear was observed22-25. The single point HSS tool specifications are as follows in figure - 2 and table - 1. Table – 1 Tool Specification Back Rake Angle 12º Side Rake Angle 12º End Relief 10º End Cutting Edge Angle 30º Side Cutting Edge Angle 15º Nose Radius 0.8mm Work piece:Work piece of standard dimensions was used for machining26-28: work piece diameter: 40mm, work piece length: 300mm (approx.). Lathe Tool Dynamometer:The instrument used for the measurement of cutting force was IEICOS multi-component force indicator. It comprises of three independent digital display calibrated to display force directly using three component tool dynamometer. This instrument comprises independent DC excitation supply for feeding strain gauge bridges, signal processing systems to process and compute respective force values for direct independent display. Instrument operates on 230V, 50Hz AC mains. To record the force readings, IEICOS multi-component force indicator software was used. The data was obtained through a USB cable connected to the Dynamometer and stored on a computer. Surface Roughness measurement:The instrument used to measure surface roughness was Surtronics 3+. For a probe movement of 4mm, surface roughness readings were recorded at three locations on the work piece and the average value was used for analysis. Specifications of Surtronics 3+: gauge range: ±150um, probe movement (max): 25.4mm, traverse speed: 1mm/s. Cutting Conditions and Experimental Procedure:Among the speed and feed rate combinations available on the Lathe, three levels of cutting parameters were selected29. It is given in table – 2. Table-2 Factors and their Levels Factor Level 1 Level 2 Level 3 A: Speed (rpm) 228 360 450 B: Feed (mm/rev) 0.11 0.18 0.75 C: Depth of Cut (mm) 0.25 0.50 0.75 Full-Factorial design for three levels and three factors (3) yielded 27 experiments and two replicates were carried out. The standard order, run order, cutting parameters and responses are as shown in the Design of Experiments table30,31. It is given in table - 3. Results and DiscussionCutting Force Analysis: Cutting force increases almost linearly with the increase in depth of cut from 0.25mm to 0.75mm. Optimum conditions are achieved for a feed rate value of 0.18 mm/rev and a DOC value of 0.5mm. Figure-4, which is the main effects plot for cutting force indicates that cutting force is influenced significantly by depth of cut, feed rate, interaction effect of feed and depth of cut and interaction effect of speed, feed and depth of cut, whereas, speed has an insignificant influence on cutting force which is shown in table - 4. Further, the model adequately explains the total variance in cutting parameters and it is also reasonably a good fit (R = 95.22%; Radj. = 90.62%). It can also be noted through ANOVA that Research Journal of Recent Sciences ______________________________________________________________ ISSN 2277-2502Vol. 1(10), 19-26, October (2012) Res. J. Recent Sci. International Science Congress Association 21 Cutting force is not significantly influenced due to the interaction between speed and feed, however, there is an indication that at higher feed rate the influence may be significant27. Figure - 5 shows the interaction plot. Also, the table - 4 shows analysis of variance. Figure-1 Experimental set-up Figure-2 HSS tool nomenclature Figure-3 IEICOS Multi-Component Force Indicator Research Journal of Recent Sciences ______________________________________________________________ ISSN 2277-2502Vol. 1(10), 19-26, October (2012) Res. J. Recent Sci. International Science Congress Association 22 Table-3 The DOE Table Standard Order Run Order Speed (rpm): A Feed (mm/rev): B DOC (mm): C Ra (µm) Fc (kgf) AB BC AC ABC 39 1 360 0.11 0.75 5.95 21 39.6 0.0825 270 29.7 5 2 228 0.18 0.5 7.34 16 41.04 0.09 114 20.52 45 3 360 0.25 0.75 7.49 32 90 0.1875 270 67.5 26 4 450 0.25 0.5 8.73 26 112.5 0.125 225 56.25 36 5 228 0.25 0.75 6.75 25 57 0.1875 171 42.75 3 6 228 0.11 0.75 4.45 20 25.08 0.0825 171 18.81 19 7 450 0.11 0.25 5.61 5 49.5 0.0275 112.5 12.375 27 8 450 0.25 0.75 9.67 30 112.5 0.1875 337.5 84.375 53 9 450 0.25 0.5 8.2 23 112.5 0.125 225 56.25 52 10 450 0.25 0.25 8.53 22 112.5 0.0625 112.5 28.125 23 11 450 0.18 0.5 7.01 14 81 0.09 225 40.5 44 12 360 0.25 0.5 10.2 23 90 0.125 180 45 31 13 228 0.18 0.25 7.86 12 41.04 0.045 57 10.26 8 14 228 0.25 0.5 10 23 57 0.125 114 28.5 20 15 450 0.11 0.5 6.01 18 49.5 0.055 225 24.75 49 16 450 0.18 0.25 7.77 12 81 0.045 112.5 20.25 7 17 228 0.25 0.25 9.6 15 57 0.0625 57 14.25 38 18 360 0.11 0.5 5.66 16 39.6 0.055 180 19.8 47 19 450 0.11 0.5 8.66 18 49.5 0.055 225 24.75 14 20 360 0.18 0.5 6.32 13 64.8 0.09 180 32.4 22 21 450 0.18 0.25 6.91 12 81 0.045 112.5 20.25 2 s22 228 0.11 0.5 3.96 15 25.08 0.055 114 12.54 21 23 450 0.11 0.75 7.36 20 49.5 0.0825 337.5 37.125 41 24 360 0.18 0.5 8.53 14 64.8 0.09 180 32.4 33 25 228 0.18 0.75 8.35 22 41.04 0.135 171 30.78 16 26 360 0.25 0.25 10.6 19 90 0.0625 90 22.5 25 27 450 0.25 0.25 9.21 15 112.5 0.0625 112.5 28.125 12 28 360 0.11 0.75 7.93 21 39.6 0.0825 270 29.7 10 29 360 0.11 0.25 4.02 12 39.6 0.0275 90 9.9 42 30 360 0.18 0.75 8.06 24 64.8 0.135 270 48.6 40 31 360 0.18 0.25 5.25 12 64.8 0.045 90 16.2 54 32 450 0.25 0.75 11.07 26 112.5 0.1875 337.5 84.375 28 33 228 0.11 0.25 4.28 10 25.08 0.0275 57 6.27 6 34 228 0.18 0.75 6.53 23 41.04 0.135 171 30.78 9 35 228 0.25 0.75 8.93 25 57 0.1875 171 42.75 43 36 360 0.25 0.25 10.93 16 90 0.0625 90 22.5 32 37 228 0.18 0.5 8.9 19 41.04 0.09 114 20.52 30 38 228 0.11 0.75 4.92 25 25.08 0.0825 171 18.81 1 39 228 0.11 0.25 4.2 8 25.08 0.0275 57 6.27 13 40 360 0.18 0.25 7.6 12 64.8 0.045 90 16.2 48 41 450 0.11 0.75 7.86 20 49.5 0.0825 337.5 37.125 17 42 360 0.25 0.5 9.53 26 90 0.125 180 45 4 43 228 0.18 0.25 9.2 12 41.04 0.045 57 10.26 18 44 360 0.25 0.75 9.8 27 90 0.1875 270 67.5 51 45 450 0.18 0.75 6.93 17 81 0.135 337.5 60.75 15 46 360 0.18 0.75 9 26 64.8 0.135 270 48.6 34 47 228 0.25 0.25 9.61 16 57 0.0625 57 14.25 24 48 450 0.18 0.75 7.59 20 81 0.135 337.5 60.75 50 49 450 0.18 0.5 6.26 12 81 0.09 225 40.5 46 50 450 0.11 0.25 6.66 6 49.5 0.0275 112.5 12.375 37 51 360 0.11 0.25 6.73 14 39.6 0.0275 90 9.9 29 52 228 0.11 0.5 5.65 15 25.08 0.055 114 12.54 35 53 228 0.25 0.5 10.0 26 57 0.125 114 28.5 11 54 360 0.11 0.5 6.86 15 39.6 0.055 180 19.8 Research Journal of Recent Sciences ______________________________________________________________ ISSN 2277-2502Vol. 1(10), 19-26, October (2012) Res. J. Recent Sci. International Science Congress Association 23 Figure-4 Main Effects Plot for Cutting Force FFigure-5 Interaction Plot for Cutting Force Fc Table-4 ANOVA for cutting force Source DF Seq SS Adj SS Adj MS F P Speed 2 20.481 20.481 10.241 2.91 0.072 Feed 2 625.815 625.815 312.907 88.93 0.000* DOC 2 1046.370 1046.370 523.185 148.69 0.000* Speed*Feed 4 36.407 36.407 9.102 2.59 0.059 Speed*DOC 4 28.519 28.519 7.130 2.03 0.119 Feed*DOC 4 50.519 50.519 12.630 3.59 0.018* Speed*Feed*DOC 8 85.259 85.259 10.657 3.03 0.015* Error 27 95.000 95.000 3.519 -- -- Total 53 1988.370 -- -- -- -- S = 1.87577 R 2 = 95.22% R 2 (adj) = 90.62% *Significant influence ( = 0.05). Research Journal of Recent Sciences ______________________________________________________________ ISSN 2277-2502Vol. 1(10), 19-26, October (2012) Res. J. Recent Sci. International Science Congress Association 24 Regression Analysis: The regression equation13 is given in table - 5: F (kgf) = 0.6 - 0.0143*speed + 34.1*feed + 29.1*depth of cut + 0.092*speed - 25*feed*depth of cut - 0.0094*speed*depth of cut + 0.005*speed*feed*depth of cut. Table-5 Regression Analysis Predictor Coef SE Coef T P Constant 0.57 13.92 0.04 0.967 Speed -0.01430 0.03890 -0.37 0.715 Feed 34.12 73.70 0.46 0.646 DOC 29.08 25.77 1.13 0.265 Speed*Feed 0.0918 0.2060 0.45 0.658 Feed*DOC -25.4 136.5 -0.19 0.853 Speed*DOC -0.00935 0.07203 -0.13 0.897 Speed*Feed*DOC 0.0046 0.3814 0.01 0.990 S = 2.98097 R-Sq = 79.4% R-Sq(adj) = 76.3% Surface Roughness Analysis: The main effects plot indicates that surface roughness is significantly influenced by feed rate and the interaction between speed and feed, which are shown in figure - 6 and 7. Optimum condition for surface roughness is achieved at a feed rate value of 0.18 mm/rev for a speed of 360 rpm and depth of cut of 0.5mm.Further, the model satisfactorily explains the total variance in cutting parameters and it is also reasonably a good fit (R = 84.25%; R adj. = 69.09%). This is shown in table - 6. Regression Analysis: The regression equation is given in table - 7: Ra (µm) = - 6.04 + 0.0187*speed + 85.4*feed + 7.42*depth of cut - 0.135*speed*feed - 69.9*feed*depth of cut - 0.0056*speed *depth of cut + 0.121*speed*feed*depth of cut. Figure-6 Main Effects Plot for Surface Roughness Ra Figure-7 Interaction Plot for Surface Roughness Ra Research Journal of Recent Sciences ______________________________________________________________ ISSN 2277-2502Vol. 1(10), 19-26, October (2012) Res. J. Recent Sci. International Science Congress Association 25 Table-6 ANOVA for Surface Roughness Source DF Seq SS Adj SS Adj MS F P Speed 2 3.497 3.497 1.748 1.63 0.215 Feed 2 107.302 107.302 53.651 49.88 0.000* DOC 2 0.516 0.516 0.258 0.24 0.788 Speed*Feed 4 19.174 19.174 4.793 4.46 0.007* Speed*DOC 4 7.230 7.230 1.808 1.68 0.184 Feed*DOC 4 6.274 6.274 1.568 1.46 0.242 Speed*Feed*DOC 8 11.368 11.368 1.421 1.32 0.275 Error 27 29.041 29.041 1.076 -- -- Total 53 184.401 -- -- -- -- S = 1.03711 R-Sq = 84.25% R-Sq(adj) = 69.09% *Significant influence ( = 0.05). Table-7 Regression Analysis Predictor Coef SE Coef T P Constant -6.043 5.074 -1.19 0.240 Speed 0.01872 0.01418 1.32 0.193 Feed 85.38 26.87 3.18 0.003 DOC 7.424 9.395 0.79 0.433 Speed*Feed -0.13516 0.07508 -1.80 0.078 Feed*DOC -69.93 49.74 -1.41 0.166 Speed*DOC -0.00561 0.02626 -0.21 0.832 Speed*Feed*DOC 0.1213 0.1390 0.87 0.387 S = 1.08666 R-Sq = 70.5% R-Sq(adj) = 66.1% ConclusionThe feed rate has significant influence on both the Cutting force and Surface roughness. Cutting Speed has no significant effect on the cutting force as well as the surface roughness of the chosen work piece. Depth of cut has a significant influence on cutting force, but an insignificant influence on surface roughness. In turning process optimization with respect to power consumption, the focus should be on choosing an appropriate combination of feed rate and depth of cut. Optimum surface roughness can be achieved by selecting relatively higher values of speed (�450 rpm), higher values of depth of cut (�0.75mm), and relatively lower values of feed rate (0.11 mm/rev). In comparison to the sequential approach adopted in most of the contemporary research, this research has shown that adjusted approach can also be successfully used to fit a reasonably acceptable and generalized model provided, it is a mono-block design. While the results declared through this experimental work may be generalized to a considerable extent while working on Mild Steel using HSS tool, the study is limited to the extreme range of values of the cutting parameters specified. Future research work may be directed towards applying Response Surface Methodology to further fine tune the optimization of cutting parameters, which was beyond the scope of this research, as it was mainly focussed towards the identification of most significantly influencing factors. Acknowledgement We acknowledge the support rendered by Manipal Institute of Technology affiliated to Manipal University, Manipal, India for providing us an opportunity to conduct experimental work on lathe, and also, supplying to us with the work piece, tool materials, measurement instruments, and all the necessary logistics as and when required. References 1.Marandet B., Verquin B., Saint-Chely J., Anderson C. and Ryckeboer M., http://aluminium.matter.org.uk (last accessed on 24th June 2011), (2011)2.Bradley C., Automated Surface Roughness Measurement, International Journal of Advanced Manufacturing Technology, 16(9), 668-674 (2000)3.Kumar P., Singh N. and Goel P., A multi-objective framework for design of vacuum-sealed molding process, Robotics and Computer Integrated Manufacturing, 15(5), 413-432 (1999) Research Journal of Recent Sciences ______________________________________________________________ ISSN 2277-2502Vol. 1(10), 19-26, October (2012) Res. J. Recent Sci. International Science Congress Association 26 4.Boothroyd G. and Knight W.A., Fundamentals of Machining and Machine Tools, 3rd Ed.,CRC Publication. ISBN 1-57444-659-2, (2006)5.Asiltürk I. and Çunkas M., Modeling and prediction of surface roughness in turning operations using artificial neural network and multiple regression method, Expert Systems with Applications, 38(5), 5826-5832 (2010)6.Davim J.P., A Note On the Determination of Optimal Cutting Conditions for Surface Finish Obtained in Turning Using Design of Experiments, Journal of Material Processing Technology, 116(2-3), 305-308 (2001)7.Meyers A.R. and Slattery T.J., Basic Machining Reference Handbook, Industrial Press Inc, (2001)8.Shaw M., Metal Cutting Principles, Oxford University Press, (1984)9.Groover M.P., Fundamentals of Modern Manufacturing, Prentice-Hall, Upper Saddle River, NJ (now published by John Wiley, New York), 634 (1996)10.Kalpakjian S. and Schmid S.R., Manufacturing Processes for Engineering Materials, 5th ed., Pearson Education, Inc., SBN 9-78013227-271-1 (2008)11.Aruna M., Dhanalakshmi V., Response surface methodology in finish turning INCONEL 718, International Journal of Engineering Science and Technology, 2(9), 4292-4297, (2010)12.Mativenga P.T., Abdukhsim N.A., Sheikh M.A. and Hon B.K.K., An Investigation of Tool Chip Contact Phenomena in High-Speed Turning Using Coated Tools, Sage Publications, 20(5), 657-667 (2006)13.Sutter G., Faure L., Molinari A., Ranc N. and Pina V., An experimental technique for the measurement of temperature fields for the orthogonal cutting in high speed machining, International Journal of Machine Tools and Manufacture, 43(7), 671-678 (2003)14.Astakhov V.P. and Shvets S.V., A Novel Approach to Operating Force Evaluation in High Strain Rate Metal-Deforming Technological process, Journal of Material Processing Technology, 117(1-2), 226-237 (2001)15.Lalwani D.I., Mehta N.K., Jain P.K., Experimental Investigations of Cutting Parameters Influence on Cutting Forces and Surface Roughness in Finish Hard Turning of MDN250 Steel, Journal of Material Processing Technology, 206(1-3), 167-179 (2007) 16.Gupta B., Gopala K.V. and Yadav J.S., Aerospace materials: With General Metallurgy for Engineers, S. Chand and Co. Ltd., New Delhi, India (1996) 17.Aneiro F.M., Coelho R.T. and Brandao L.C., Turning Hardened Steel Using Coated Carbide at High Cutting Speeds, Journal of the Brazilian Society of Mechanical Sciences and Engineering, 30(2), 104-109 (2008)18.Abele E., Dietz S. and Schiffler A., Analysis of Cutting Force during Milling with regards to the dependency on the penetration angle, Production Engineering, , 483-487 (2009)19.Patel K., Batish A. and Bhattacharya A., Optimization of Surface Roughness in an end-milling operation using nested experimental design, Production Engineering, 3(4-5), 361-373 (2009)20.Miles J., R-squared, Adjusted R-squared, Encyclopaedia of Statisticsin Behavioural Science, John Wiley and Sons, Ltd, (2005)21.Ricci L. and Martínez R., Adjusted R2-type measures for Tweedie models, Computational Statistics and Data Analysis, 52(3), 1650-1660 (2008)22.Sharma V.S., Sharma S.K. and Sharma A.K., Cutting Tool Wear Estimation for Turning, Journal of Intelligent Manufacturing,19(1), 99-108 (2007)23.Tsao C.C., An experiment study of hard coating and cutting fluid effect in milling aluminum alloy, International Journal of Advanced Manufacturing Technology, 32(9-10), 885-891 (2006)24.Wang J., Liu Y.B., An J. and Wang L.M., Wear mechanism map of uncoated HSS tools during drilling die-cast magnesium alloy, Wear, 265(5-6), 685-691 (2008)25.Xie L.J., Schmidt J., Schmidt C. and Biesinger F., 2D FEM estimate of tool wear in turning operation, Wear, 258(10), 1479-1450 (2005)26.Oishi K., Built-up edge elimination in mirror cutting of hardened steel, International Journal of Machine Tools Manufacture, 39, 885-903 (1995) 27.Yang W. and Tarng Y., Design optimization of cutting parameters for turning operations, Journal of Material Processing Technology, 84(1-3), 122-129 (1998)28.Davim P., Influence of cutting parameters of surface finish obtained by turning, M.Sc. Thesis, University of Porto, Porto, 10-128 (1990)29.Lu C., Ma N., Chen Z. and Costes J.P., Pre-evaluation on surface profile in turning process based on cutting parameters, International Journal of Advanced Manufacturing Technology,49(5-8), 447-458 (2010)30.Montgomery D.C., Design and Analysis of Experiments, 5th ed. John Wiley and Sons Inc., (2001)31.Das M.N. and Giri N.C., Design and Analysis of Experiments, Second ed., New Age International (P) Ltd., New Delhi, India. ISBN: 0-85226-914-5 (1999)