Towards Optimization of Surface Roughness and Productivity Aspects during High-Speed Machining of Ti–6Al–4V
Abstract
:1. Introduction
2. Experimentation and Optimization Approach
- Step 1: Initially, the input parameters and performance characteristics were defined.
- Step 2: The data should be transformed into a matrix form (decision matrix) as provided by Equation (1):
- Step 3: The ratio can be investigated using Equation (2). The square root of the sum of square for each alternative is to be the best choice for the denominator.
- Step 1: In the first step, the data obtained from the experiments were analyzed.
- Step 2: In the second step, the response characteristics were normalized to investigate the performance indices using MOORA.
- Step 3: The statistical analysis of the performance index was made by MiniTab statistical software.
- Step 4: In the next step, the regression coefficients with empirical models were investigated. The empirical model of Pi varies the Pi with the input process variables.
- Step 5: In the fifth step, the empirical model was solved varying the population and generation using PSO. With the implementation of PSO, the global best velocity and positions were determined.
- Step 6: Finally, confirmation experiments were performed at the machining setting suggested by the MCDM approach to analyze the performance characteristics of machining for titanium.
3. Results and Discussions: Analysis, Optimization and Validation
0.1 ≤ DoC ≤ 0.3
0.05 ≤ Feed≤ 0.15
5 ≤ Cutting Length ≤ 120
Performance Index = 0.12 − 0.0005 × speed − 0.79 × DoC + 3.218 × F + 0.0012 × Cutting Length + 0.0026 × speed × DoC + 0.0046 × speed × F
- Minimizing Ra and Rq
- Maximizing MRR
- Multi-objective optimization between Ra, Rq and MRR
4. Conclusions and Future Work
Author Contributions
Funding
Acknowledgments
Conflicts of Interest
References
- Jamil, M.; Khan, A.M.; Hegab, H.; Gong, L.; Mia, M.; Gupta, M.K.; He, N. Effects of hybrid Al2O3-CNT nanofluids and cryogenic cooling on machining of Ti–6Al–4V. Int. J. Adv. Manuf. Technol. 2019, 102, 3895–3909. [Google Scholar] [CrossRef]
- Mia, M.; Dhar, N.R. Effects of duplex jets high-pressure coolant on machining temperature and machinability of Ti-6Al-4V superalloy. J. Mater. Process. Technol. 2018, 252, 688–696. [Google Scholar] [CrossRef]
- Arrazola, P.J.; Garay, A.; Iriarte, L.M.; Armendia, M.; Marya, S.; Le Maître, F. Machinability of titanium alloys (Ti6Al4V and Ti555.3). J. Mater. Process. Technol. 2009, 209, 2223–2230. [Google Scholar] [CrossRef]
- Komanduri, R. Some clarifications on the mechanics of chip formation when machining titanium alloys. Wear 1982, 76, 15–34. [Google Scholar] [CrossRef]
- Sun, J.; Guo, Y. A comprehensive experimental study on surface integrity by end milling Ti–6Al–4V. J. Mater. Process. Technol. 2009, 209, 4036–4042. [Google Scholar] [CrossRef]
- Ginting, A.; Nouari, M. Surface integrity of dry machined titanium alloys. Int. J. Mach. Tools Manuf. 2009, 49, 325–332. [Google Scholar] [CrossRef]
- Selvakumar, S.; Ravikumar, R.; Raja, K. Implementation of response surface methodology in finish turning on titanium alloy Gr. 2. Eur. J. Sci. Res. 2012, 81, 436–445. [Google Scholar]
- Pervaiz, S.; Deiab, I.; Rashid, A.; Nicolescu, M.; Kishawy, H. Energy consumption and surface finish analysis of machining Ti6Al4V. World Acad. Sci. Eng. Technol. 2013, 76, 113–118. [Google Scholar]
- Fang, N.; Wu, Q. A comparative study of the cutting forces in high speed machining of Ti–6Al–4V and Inconel 718 with a round cutting edge tool. J. Mater. Process. Technol. 2009, 209, 4385–4389. [Google Scholar] [CrossRef]
- D’Mello, G.; Pai, P.S.; Puneet, N.; Fang, N. Surface roughness evaluation using cutting vibrations in high speed turning of Ti-6Al-4V-an experimental approach. Int. J. Mach. Mach. Mater. 2016, 18, 288–312. [Google Scholar]
- Ezugwu, E.O.; Bonney, J.; Da Silva, R.B.; Cakir, O. Surface integrity of finished turned Ti–6Al–4V alloy with PCD tools using conventional and high pressure coolant supplies. Int. J. Mach. Tools Manuf. 2007, 47, 884–891. [Google Scholar] [CrossRef]
- Yigit, R.; Celik, E.; Findik, F.; Koksal, S. Effect of cutting speed on the performance of coated and uncoated cutting tools in turning nodular cast iron. J. Mater. Process. Technol. 2008, 204, 80–88. [Google Scholar] [CrossRef]
- Haron, C.C.; Ginting, A.; Goh, J. Wear of coated and uncoated carbides in turning tool steel. J. Mater. Process. Technol. 2001, 116, 49–54. [Google Scholar] [CrossRef]
- Sun, S.; Brandt, M.; Mo, J.P. Evolution of tool wear and its effect on cutting forces during dry machining of Ti-6Al-4V alloy. Proc. Inst. Mech. Eng. Part B J. Eng. Manuf. 2014, 228, 191–202. [Google Scholar] [CrossRef]
- Ezugwu, E.; Da Silva, R.; Bonney, J.; Machado, A. Evaluation of the performance of CBN tools when turning Ti–6Al–4V alloy with high pressure coolant supplies. Int. J. Mach. Tools Manuf. 2005, 45, 1009–1014. [Google Scholar] [CrossRef]
- Mello, G.D.; Pai, P.S. Surface Roughness Modeling in High Speed Turning of Ti-6Al-4V Using Response Surface Methodology. Mater. Today Proc. 2018, 5, 11686–11696. [Google Scholar] [CrossRef]
- Ramesh, S.; Karunamoorthy, L.; Palanikumar, K. Measurement and analysis of surface roughness in turning of aerospace titanium alloy (gr5). Measurement 2012, 45, 1266–1276. [Google Scholar] [CrossRef]
- Sen, B.; Hussain, S.A.I.; Mia, M.; Mandal, U.K.; Mondal, S.P. Selection of an ideal MQL-assisted milling condition: An NSGA-II-coupled TOPSIS approach for improving machinability of Inconel 690. Int. J. Adv. Manuf. Technol. 2019, 103, 1811–1829. [Google Scholar] [CrossRef]
- Bhowmik, S.; Paul, A.; Panua, R.; Ghosh, S.K.; Debroy, D. Performance-exhaust emission prediction of diesosenol fueled diesel engine: An ANN coupled MORSM based optimization. Energy 2018, 153, 212–222. [Google Scholar] [CrossRef]
- Ranganathan, S.; Senthilvelan, T.; Sriram, G. Evaluation of machining parameters of hot turning of stainless steel (Type 316) by applying ANN and RSM. Mater. Manuf. Process. 2010, 25, 1131–1141. [Google Scholar] [CrossRef]
- Brauers, W.K. Optimization Methods for a Stakeholder Society: A Revolution in Economic Thinking by Multi-Objective Optimization; Springer Science & Business Media: Berlin/Heidelberg, Germany, 2013; Volume 73. [Google Scholar]
- Abbas, A.T.; Benyahia, F.; El Rayes, M.M.; Pruncu, C.; Taha, M.A.; Hegab, H. Towards optimization of machining performance and sustainability aspects when turning AISI 1045 Steel under different cooling and lubrication strategies. Materials 2019, 12, 3023. [Google Scholar] [CrossRef] [PubMed]
- Hegab, H.; Kishawy, H.A.; Umer, U.; Mohany, A. A model for machining with nano-additives based minimum quantity lubrication. Int. J. Adv. Manuf. Technol. 2019, 102, 1–16. [Google Scholar] [CrossRef]
N | C | H | Fe | O | Al | V | Ti |
---|---|---|---|---|---|---|---|
0.05% | 0.1% | 0.012% | 0.4% | 0.2% | 6% | 4% | Balance |
Tensile Strength (MPa) | Yield Strength, 0.2% Offset (MPa) | Elongation (%) | Reduction of Area (%) | Hardness |
---|---|---|---|---|
895 | 825 | 10 | 25 | HRC 36 |
Test # | Speed (m/min) | Depth of Cut (mm) | Feed (mm/rev) | Cutting Length (mm) | Ra (µm) | Rq (μm) | MRR (mm3/min) |
---|---|---|---|---|---|---|---|
1 | 300 | 0.1 | 0.05 | 5 | 0.590 | 0.707 | 1500 |
2 | 300 | 0.1 | 0.05 | 40 | 0.767 | 0.866 | 1500 |
3 | 300 | 0.1 | 0.05 | 80 | 1.048 | 1.210 | 1500 |
4 | 300 | 0.1 | 0.05 | 120 | 1.183 | 1.335 | 1500 |
5 | 300 | 0.1 | 0.15 | 5 | 2.001 | 2.459 | 4500 |
6 | 300 | 0.1 | 0.15 | 40 | 2.539 | 3.062 | 4500 |
7 | 300 | 0.1 | 0.15 | 80 | 2.917 | 3.553 | 4500 |
8 | 300 | 0.1 | 0.15 | 120 | 4.326 | 5.213 | 4500 |
9 | 200 | 0.1 | 0.05 | 5 | 0.470 | 0.598 | 1000 |
10 | 200 | 0.1 | 0.05 | 40 | 0.486 | 0.603 | 1000 |
11 | 200 | 0.1 | 0.05 | 80 | 0.601 | 0.758 | 1000 |
12 | 200 | 0.1 | 0.05 | 120 | 0.689 | 0.916 | 1000 |
13 | 200 | 0.1 | 0.15 | 5 | 1.497 | 1.762 | 3000 |
14 | 200 | 0.1 | 0.15 | 40 | 1.539 | 1.883 | 3000 |
15 | 200 | 0.1 | 0.15 | 80 | 1.808 | 2.064 | 3000 |
16 | 200 | 0.1 | 0.15 | 120 | 2.262 | 2.520 | 3000 |
17 | 100 | 0.1 | 0.05 | 5 | 0.244 | 0.310 | 500 |
18 | 100 | 0.1 | 0.05 | 40 | 0.295 | 0.392 | 500 |
19 | 100 | 0.1 | 0.05 | 80 | 0.302 | 0.384 | 500 |
20 | 100 | 0.1 | 0.05 | 120 | 0.370 | 0.455 | 500 |
21 | 100 | 0.1 | 0.15 | 5 | 1.747 | 2.264 | 1500 |
22 | 100 | 0.1 | 0.15 | 40 | 1.869 | 2.254 | 1500 |
23 | 100 | 0.1 | 0.15 | 80 | 1.910 | 2.447 | 1500 |
24 | 100 | 0.1 | 0.15 | 120 | 2.146 | 2.502 | 1500 |
25 | 300 | 0.3 | 0.05 | 5 | 0.687 | 0.813 | 4500 |
26 | 300 | 0.3 | 0.05 | 40 | 0.722 | 0.818 | 4500 |
27 | 300 | 0.3 | 0.05 | 80 | 0.810 | 0.935 | 4500 |
28 | 300 | 0.3 | 0.05 | 120 | 1.031 | 1.212 | 4500 |
29 | 300 | 0.3 | 0.15 | 5 | 1.701 | 1.997 | 13,500 |
30 | 300 | 0.3 | 0.15 | 40 | 1.877 | 2.213 | 13,500 |
31 | 300 | 0.3 | 0.15 | 80 | 4.700 | 5.352 | 13,500 |
32 | 300 | 0.3 | 0.15 | 120 | 4.956 | 6.240 | 13,500 |
33 | 200 | 0.3 | 0.05 | 5 | 1.008 | 1.482 | 3000 |
34 | 200 | 0.3 | 0.05 | 40 | 1.104 | 1.529 | 3000 |
35 | 200 | 0.3 | 0.05 | 80 | 1.430 | 1.728 | 3000 |
36 | 200 | 0.3 | 0.05 | 120 | 1.533 | 1.901 | 3000 |
37 | 200 | 0.3 | 0.15 | 5 | 1.821 | 2.300 | 9000 |
38 | 200 | 0.3 | 0.15 | 40 | 1.891 | 2.355 | 9000 |
39 | 200 | 0.3 | 0.15 | 80 | 2.002 | 2.235 | 9000 |
40 | 200 | 0.3 | 0.15 | 120 | 2.576 | 2.808 | 9000 |
41 | 100 | 0.3 | 0.05 | 5 | 0.440 | 0.533 | 1500 |
42 | 100 | 0.3 | 0.05 | 40 | 0.633 | 0.761 | 1500 |
43 | 100 | 0.3 | 0.05 | 80 | 0.923 | 1.183 | 1500 |
44 | 100 | 0.3 | 0.05 | 120 | 0.993 | 1.218 | 1500 |
45 | 100 | 0.3 | 0.15 | 5 | 1.452 | 1.802 | 4500 |
46 | 100 | 0.3 | 0.15 | 40 | 1.633 | 1.911 | 4500 |
47 | 100 | 0.3 | 0.15 | 80 | 1.727 | 1.999 | 4500 |
48 | 100 | 0.3 | 0.15 | 120 | 1.788 | 2.110 | 4500 |
Test # | Ra2 | Rq2 | MRR2 | Normalized Ra | Normalized Rq | Normalized MRR | Performance Index (Pi) |
---|---|---|---|---|---|---|---|
1 | 0.3481 | 0.4998 | 2,250,000 | 0.0461 | 0.0460 | 0.0401 | 0.1322 |
2 | 0.5883 | 0.7500 | 2,250,000 | 0.0599 | 0.0563 | 0.0401 | 0.1563 |
3 | 1.0983 | 1.4641 | 2,250,000 | 0.0819 | 0.0787 | 0.0401 | 0.2007 |
4 | 1.3995 | 1.7822 | 2,250,000 | 0.0924 | 0.0869 | 0.0401 | 0.2193 |
5 | 4.0040 | 6.0467 | 20,250,000 | 0.1563 | 0.1600 | 0.1203 | 0.4365 |
6 | 6.4465 | 9.3758 | 20,250,000 | 0.1983 | 0.1992 | 0.1203 | 0.5178 |
7 | 8.5089 | 12.6238 | 20,250,000 | 0.2278 | 0.2312 | 0.1203 | 0.5793 |
8 | 18.7143 | 27.1754 | 20,250,000 | 0.3379 | 0.3392 | 0.1203 | 0.7973 |
9 | 0.2209 | 0.3576 | 1,000,000 | 0.0367 | 0.0389 | 0.0267 | 0.1023 |
10 | 0.2362 | 0.3636 | 1,000,000 | 0.0380 | 0.0392 | 0.0267 | 0.1039 |
11 | 0.3612 | 0.5746 | 1,000,000 | 0.0469 | 0.0493 | 0.0267 | 0.1230 |
12 | 0.4747 | 0.8391 | 1,000,000 | 0.0538 | 0.0596 | 0.0267 | 0.1401 |
13 | 2.2410 | 3.1046 | 9,000,000 | 0.1169 | 0.1146 | 0.0802 | 0.3117 |
14 | 2.3685 | 3.5457 | 9,000,000 | 0.1202 | 0.1225 | 0.0802 | 0.3229 |
15 | 3.2689 | 4.2601 | 9,000,000 | 0.1412 | 0.1343 | 0.0802 | 0.3557 |
16 | 5.1166 | 6.3504 | 9,000,000 | 0.1767 | 0.1639 | 0.0802 | 0.4208 |
17 | 0.0595 | 0.0961 | 250,000 | 0.0191 | 0.0202 | 0.0134 | 0.0526 |
18 | 0.0870 | 0.1537 | 250,000 | 0.0230 | 0.0255 | 0.0134 | 0.0619 |
19 | 0.0912 | 0.1475 | 250,000 | 0.0236 | 0.0250 | 0.0134 | 0.0619 |
20 | 0.1369 | 0.2070 | 250,000 | 0.0289 | 0.0296 | 0.0134 | 0.0719 |
21 | 3.0520 | 5.1257 | 2,250,000 | 0.1365 | 0.1473 | 0.0401 | 0.3238 |
22 | 3.4932 | 5.0805 | 2,250,000 | 0.1460 | 0.1466 | 0.0401 | 0.3327 |
23 | 3.6481 | 5.9878 | 2,250,000 | 0.1492 | 0.1592 | 0.0401 | 0.3485 |
24 | 4.6053 | 6.2600 | 2,250,000 | 0.1676 | 0.1628 | 0.0401 | 0.3705 |
25 | 0.4720 | 0.6610 | 20,250,000 | 0.0537 | 0.0529 | 0.1203 | 0.2268 |
26 | 0.5213 | 0.6691 | 20,250,000 | 0.0564 | 0.0532 | 0.1203 | 0.2299 |
27 | 0.6561 | 0.8742 | 20,250,000 | 0.0633 | 0.0608 | 0.1203 | 0.2444 |
28 | 1.0630 | 1.4689 | 20,250,000 | 0.0805 | 0.0789 | 0.1203 | 0.2797 |
29 | 2.8934 | 3.9880 | 182,250,000 | 0.1329 | 0.1299 | 0.3608 | 0.6236 |
30 | 3.5231 | 4.8974 | 182,250,000 | 0.1466 | 0.1440 | 0.3608 | 0.6514 |
31 | 22.0900 | 28.6439 | 182,250,000 | 0.3671 | 0.3482 | 0.3608 | 1.0761 |
32 | 24.5619 | 38.9376 | 182,250,000 | 0.3871 | 0.4060 | 0.3608 | 1.1539 |
33 | 1.0161 | 2.1963 | 9,000,000 | 0.0787 | 0.0964 | 0.0802 | 0.2553 |
34 | 1.2188 | 2.3378 | 9,000,000 | 0.0862 | 0.0995 | 0.0802 | 0.2659 |
35 | 2.0449 | 2.9860 | 9,000,000 | 0.1117 | 0.1124 | 0.0802 | 0.3043 |
36 | 2.3501 | 3.6138 | 9,000,000 | 0.1197 | 0.1237 | 0.0802 | 0.3236 |
37 | 3.3160 | 5.2900 | 81,000,000 | 0.1422 | 0.1496 | 0.2405 | 0.5324 |
38 | 3.5759 | 5.5460 | 81,000,000 | 0.1477 | 0.1532 | 0.2405 | 0.5415 |
39 | 4.0080 | 4.9952 | 81,000,000 | 0.1564 | 0.1454 | 0.2405 | 0.5423 |
40 | 6.6358 | 7.8849 | 81,000,000 | 0.2012 | 0.1827 | 0.2405 | 0.6244 |
41 | 0.1936 | 0.2841 | 2,250,000 | 0.0344 | 0.0347 | 0.0401 | 0.1091 |
42 | 0.4007 | 0.5791 | 2,250,000 | 0.0494 | 0.0495 | 0.0401 | 0.1390 |
43 | 0.8519 | 1.3995 | 2,250,000 | 0.0721 | 0.0770 | 0.0401 | 0.1891 |
44 | 0.9860 | 1.4835 | 2,250,000 | 0.0776 | 0.0792 | 0.0401 | 0.1969 |
45 | 2.1083 | 3.2472 | 20,250,000 | 0.1134 | 0.1172 | 0.1203 | 0.3509 |
46 | 2.6667 | 3.6519 | 20,250,000 | 0.1276 | 0.1243 | 0.1203 | 0.3721 |
47 | 2.9825 | 3.9960 | 20,250,000 | 0.1349 | 0.1301 | 0.1203 | 0.3852 |
48 | 3.1969 | 4.4521 | 20,250,000 | 0.1397 | 0.1373 | 0.1203 | 0.3972 |
Source | Degree of Freedom (DF) | Statistical Summation (SS) | Contribution Percentage (%) | Mean Square (MS) | F-Value | P-Value |
---|---|---|---|---|---|---|
Model | 11 | 2.34519 | 0.21320 | 20.98 | 0.000 | |
Speed | 2 | 0.04423 | 1.63 | 0.02211 | 2.18 | 0.128 |
DoC* | 1 | 0.00427 | 0.16 | 0.00427 | 0.42 | 0.521 |
F* | 1 | 1.39348 | 51.4 | 1.39348 | 137.11 | 0.000 |
Cutting Length | 3 | 0.12240 | 4.51 | 0.04080 | 4.01 | 0.015 |
Speed*DoC | 2 | 0.68923 | 25.42 | 0.34462 | 33.91 | 0.000 |
Speed*F | 2 | 0.09158 | 3.38 | 0.04579 | 4.51 | 0.018 |
Error | 36 | 0.36586 | 13.5 | 0.01016 | ||
Total | 47 | 2.71106 |
Cutting Conditions | Initial Machining Parameter | Predicted | Experimental | |
---|---|---|---|---|
Ra and Rq | MRR | |||
(S)100(DoC)0.1(F)0.05(CL)5 | (S)300(DoC)0.3(F)0.15(CL)120 | (S)190(DoC)0.1(F)0.15(CL)120 | (S)190(DoC)0.1(F)0.15(CL)120 | |
Ra | 0.244 | 4.956 | 2.262 | 2.19 |
Rq | 0.31 | 6.240 | 2.52 | 2.43 |
MRR | 500 | 13,500 | 3000 | 2850 |
Pi | - | - | 0.6074 | - |
© 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).
Share and Cite
Abbas, A.T.; Sharma, N.; Anwar, S.; Hashmi, F.H.; Jamil, M.; Hegab, H. Towards Optimization of Surface Roughness and Productivity Aspects during High-Speed Machining of Ti–6Al–4V. Materials 2019, 12, 3749. https://doi.org/10.3390/ma12223749
Abbas AT, Sharma N, Anwar S, Hashmi FH, Jamil M, Hegab H. Towards Optimization of Surface Roughness and Productivity Aspects during High-Speed Machining of Ti–6Al–4V. Materials. 2019; 12(22):3749. https://doi.org/10.3390/ma12223749
Chicago/Turabian StyleAbbas, Adel T., Neeraj Sharma, Saqib Anwar, Faraz H. Hashmi, Muhammad Jamil, and Hussien Hegab. 2019. "Towards Optimization of Surface Roughness and Productivity Aspects during High-Speed Machining of Ti–6Al–4V" Materials 12, no. 22: 3749. https://doi.org/10.3390/ma12223749
APA StyleAbbas, A. T., Sharma, N., Anwar, S., Hashmi, F. H., Jamil, M., & Hegab, H. (2019). Towards Optimization of Surface Roughness and Productivity Aspects during High-Speed Machining of Ti–6Al–4V. Materials, 12(22), 3749. https://doi.org/10.3390/ma12223749