Tool Wear
In this study, when the maximum flank wear reached 0.2 mm, it is assumed that the cutting tool had completed its life. Tool life experiments were carried out at 80 m/min cutting speed and 0.1 mm/rev feed rate. Figure 3 shows the change in tool life depending on the machining parameters and cooling conditions. In Fig. 3, the longest tool life was obtained with MQL in experiments, while the shortest tool life value obtained from dry cutting condition. It is attributed that lubrication forms a thin layer and therefore positively affects tool life. The similar results can be found in the literature [8, 31–34].
One of the most important factors for tool life is the temperature that occurs during cutting. Increased cutting temperature may cause the tool to wears faster and thus completes its life faster. The film layer formed with MQL and the decrease of the cutting temperature provided the best result in terms of tool life. Compressed air cooling also caused the cutting zone temperature to decrease. As a result, the second best tool life after MQL was obtained in experiments using with vortex.
In this study, all tools were cut four times for 1.5 minutes (6 minutes in total) with a cutting speed of 80 m/min and a feed rate of 0.1 mm/rev Tool wear according to machining time in different cutting conditions are given in Fig. 4.
In Fig. 4, initially the wear values are very close to each other, besides the best result obtained in dry condition at 1.5 minute machining time. However, with the prolonged cutting time, obvious differences have started to occur on tool wear. Increased cutting temperatures along with increasing cutting time caused the tool to wear more in dry condition. In the experiments carried out under MQL and Vortex conditions, cutting temperatures were lower than the dry cutting condition. The minimum quantity lubrication method covers the tool-chip interface with a thin oil film, reducing friction and preventing heat build-up.
To prevent or minimize tool wear, the mechanisms that caused wear needed to be analyzed. In this study, SEM images of cutting tools were taken after the experiments and EDX analyzes were performed. SEM images of cutting tools and EDX analysis results are given in Figs. 5–7.
One of the factors affecting tool wear is Built up edge (BUE) formation. According to EDX analyzes, it is seen that alloying elements such as Ni and Cr in the composition of the workpiece material are densely encountered. BUE formation was often seen in the processing of difficult-to-cut materials such as nickel and cobalt-based superalloys. It is seen in the Figs. 5–7 where BUE formation was observed on flank side of the tool in dry condition. However, BUE formation was observed on rake in Vortex. EDX analyzes of these zones also support this situation.
Evaluation of cutting zone temperature
During machining, keeping the temperature under control has a great importance. Uncontrollable temperature plays a crucial role especially in terms of tool life. In addition to that, it significantly affects the surface quality and surface integrity. In this part of the work the effects of cutting parameters and cutting conditions on cutting zone temperature were investigated. In Fig. 8, it is seen that the highest cutting zone temperature values are reached under dry cutting condition. It is possible to say that the lowest cutting zone temperature value occurred in experiments using MQL. MQL reduces friction at the tool chip interface and also reduces cutting zone temperature values. It has been observed that the temperature values obtained with vortex are close to MQL and lower than dry conditions. It is possible to say that these results are similar to the previous studies in literature. Ji et al. compared cooling conditions (dry, MQL and flood cooling methods) and concluded that minimum quantity lubrication results lower temperature values than dry conditions [34–36]. In general, it is possible to mention that the cutting zone temperature decreases with increasing chip removal during machining. However, with the increasing feed rates and cutting speeds, the temperature values decreased contrarily to the expectation. A large amount of the heat that occur occured during machining is removed by chip. This rate can reach up to 80% under ideal conditions. But in some cases it may decrease up to 50%. It is thought that the ideal rate here occurs at 0.12 mm/rev. and 100 m/min. In all cutting conditions, this has been the case. The highest cutting temperature occurred in dry condition at 60 m/min cutting speed and 0.08 mm/rev feed rate. The lowest cutting temperature occurred in experiments with MQL at 100 m/min cutting speed and 0.12 mm/rev feed rate.
Evaluation of surface roughness
In Fig. 9, it can be seen that the surface quality improves with increasing cutting speed. This is the case for all three cutting conditions. The best surface quality occurred in experiments with MQL, followed by vortex. Improvement in surface roughness was expected with lubrication and cooling. It is possible to say that lubrication was more effective than other conditions. The lubricant that penetrates the tool chip interface facilitates the cutting and positively affects the surface quality [37]. In addition, lower temperatures and forces in MQL and vortex compared to dry condition affected the surface quality positively. Another reason for this improvement is that the chips can be removed more easily than dry condition. The chip that cannot be removed during cutting affects the surface quality negatively. In all three cutting conditions, the lowest surface roughness value was reached at 100 m/min cutting speed and 0.08 mm/rev feed rate. Even though the lowest surface roughness value was obtained in this parameter in all three cutting conditions, there was a 30% difference between MQL and dry, and a 29% difference between Vortex and dry.
Evaluation of cutting forces
In Fig. 10, it is seen that increasing cutting speeds decrease the cutting forces. It is a known situation that the temperature in the cutting zone increases with increasing cutting speeds and therefore the cutting process becomes easier. In the section where the cutting zone temperature was evaluated, it was seen that the temperature decreased with increasing cutting speeds. But here the decrease in cutting temperature has been attributed to the higher temperature removed by the chip. The increase in cutting forces with increasing feed rates can be attributed to the increase in the amount of chip removed per unit time. As with all other outputs, the best result in terms of cutting forces was obtained in experiments with MQL. It forms a film layer with the lubricating effect of MQL. As a result, decrease in cutting forces is a situation encountered in the literature [38–39]. The lowest cutting force was obtained in the experiments using MQL at 100 m/min cutting speed and 0.08 mm/rev feed rate. Then, it was carried out in experiments using Vortex at 100 m/min cutting speed and 0.08 mm/rev feed rate. The highest cutting force occurred in dry condition at 60 m/min cutting speed and 0.12 mm/rev feed rate.
Multiple Optimization
In the second phase of this study, the cooling conditions and cutting parameters were optimized according to the results of T, Fc and Ra in turning of Inconel 625 superalloy. In studies based on experimental design and analysis, especially in different cooling conditions and cutting parameters, each result is very effective and important to a certain extent, so all results are optimized at the same time (Table 4).
For this reason, Taguchi based grey relational analysis methodology was used to improve and optimize the parameters affecting the results. In the present study, the "smaller is better" approach was applied to minimize T, Fc and Ra simultaneously in the multiple response optimization process. First, the experimental results were normalized and the S/N values of the multiple response were obtained [27–30]. The values obtained as a result of the experiments and calculations are given in Table 4. In this table, the high GRG value indicates the optimum level with the strong relationship between the experimental results and the normalized values. Also, the response table for the GRG is given in Table 5. The maximum value corresponding to each parameter in this table indicates the optimum level. From now on, the optimal parameter level can be determined using Fig. 11 and/or the response Table 5. Accordingly, the best machining parameters were determined as 100 m/min cutting speed and 0.08 mm/rev feed rate in MQL cooling.
Table 4
Results of grey relational analysis
Exp
no
|
Experiment results
|
Normalized values
|
Coefficients
|
|
|
|
Fc
|
T
|
Ra
|
Fc
|
T
|
Ra
|
Fc
|
T
|
Ra
|
GRG
|
S/N
|
Order
|
1
|
351
|
491
|
1.064
|
0.537
|
0.000
|
0.891
|
0.519
|
0.333
|
0.821
|
1.674
|
-4.47639
|
15
|
2
|
397
|
458
|
1.453
|
0.224
|
0.185
|
0.803
|
0.392
|
0.380
|
0.718
|
1.490
|
-3.46442
|
21
|
3
|
430
|
360
|
4.383
|
0.000
|
0.736
|
0.141
|
0.333
|
0.654
|
0.368
|
1.356
|
-2.64314
|
24
|
4
|
321
|
472
|
0.756
|
0.741
|
0.107
|
0.961
|
0.659
|
0.359
|
0.928
|
1.946
|
-5.78246
|
10
|
5
|
360
|
438
|
0.844
|
0.476
|
0.298
|
0.941
|
0.488
|
0.416
|
0.895
|
1.799
|
-5.10098
|
14
|
6
|
411
|
361
|
0.921
|
0.129
|
0.730
|
0.924
|
0.365
|
0.650
|
0.868
|
1.882
|
-5.49323
|
12
|
7
|
295
|
462
|
0.763
|
0.918
|
0.163
|
0.959
|
0.860
|
0.374
|
0.925
|
2.159
|
-6.68345
|
7
|
8
|
350
|
377
|
0.796
|
0.544
|
0.640
|
0.952
|
0.523
|
0.582
|
0.912
|
2.017
|
-6.09461
|
8
|
9
|
402
|
351
|
0.980
|
0.190
|
0.787
|
0.910
|
0.382
|
0.701
|
0.848
|
1.931
|
-5.71357
|
11
|
10
|
324
|
478
|
0.634
|
0.721
|
0.073
|
0.989
|
0.642
|
0.350
|
0.978
|
1.970
|
-5.88961
|
9
|
11
|
354
|
405
|
0.928
|
0.517
|
0.483
|
0.922
|
0.509
|
0.492
|
0.865
|
1.866
|
-5.41663
|
13
|
12
|
395
|
391
|
1.432
|
0.238
|
0.562
|
0.808
|
0.396
|
0.533
|
0.723
|
1.652
|
-4.35948
|
16
|
13
|
313
|
416
|
4.580
|
0.796
|
0.421
|
0.097
|
0.710
|
0.464
|
0.356
|
1.530
|
-3.69363
|
20
|
14
|
345
|
377
|
4.650
|
0.578
|
0.640
|
0.081
|
0.542
|
0.582
|
0.352
|
1.476
|
-3.38443
|
22
|
15
|
371
|
340
|
5.007
|
0.401
|
0.848
|
0.000
|
0.455
|
0.767
|
0.333
|
1.556
|
-3.83842
|
18
|
16
|
283
|
392
|
0.584
|
1.000
|
0.556
|
1.000
|
1.000
|
0.530
|
1.000
|
2.530
|
-8.06159
|
1
|
17
|
328
|
356
|
0.656
|
0.694
|
0.758
|
0.984
|
0.620
|
0.674
|
0.968
|
2.263
|
-7.09249
|
5
|
18
|
371
|
313
|
0.690
|
0.401
|
1.000
|
0.976
|
0.455
|
1.000
|
0.954
|
2.409
|
-7.63809
|
2
|
19
|
340
|
482
|
3.788
|
0.612
|
0.051
|
0.276
|
0.563
|
0.345
|
0.408
|
1.317
|
-2.38879
|
25
|
20
|
380
|
411
|
4.245
|
0.340
|
0.449
|
0.172
|
0.431
|
0.476
|
0.377
|
1.284
|
-2.16888
|
27
|
21
|
413
|
375
|
4.737
|
0.116
|
0.652
|
0.061
|
0.361
|
0.589
|
0.347
|
1.298
|
-2.26599
|
26
|
22
|
315
|
425
|
3.444
|
0.782
|
0.371
|
0.353
|
0.697
|
0.443
|
0.436
|
1.576
|
-3.94860
|
17
|
23
|
350
|
407
|
4.229
|
0.544
|
0.472
|
0.176
|
0.523
|
0.486
|
0.378
|
1.387
|
-2.84215
|
23
|
24
|
400
|
338
|
4.309
|
0.204
|
0.860
|
0.158
|
0.386
|
0.781
|
0.373
|
1.539
|
-3.74528
|
19
|
25
|
292
|
420
|
0.591
|
0.939
|
0.399
|
0.998
|
0.891
|
0.454
|
0.997
|
2.342
|
-7.39113
|
3
|
26
|
324
|
356
|
0.688
|
0.721
|
0.758
|
0.976
|
0.642
|
0.674
|
0.955
|
2.271
|
-7.12532
|
4
|
27
|
377
|
326
|
0.840
|
0.361
|
0.927
|
0.942
|
0.439
|
0.873
|
0.896
|
2.208
|
-6.87849
|
6
|
Table 5
Response table for the GRG
Parameters
|
Level 1
|
Level 2
|
Level 3
|
Delta
|
Cooling conditions
|
1.806
|
1.917
|
1.691
|
0.226
|
Cutting speed
|
1.545
|
1.632
|
2.237
|
0.691
|
Feed rate
|
1.894
|
1.761
|
1.759
|
0.135
|
Total average value of the GRG = 2.015
|
Verification of Optimization
The last step in Taguchi based grey relational analysis is the verification of the optimum parameter determined. Verification experiments were performed three times using the determined optimum parameters and the results are given in Table 6.
Table 6
Results of the confirmation test
Initial parameters
|
Optimum parameters
|
Prediction
|
Experiment
|
Level
|
A1-B2-C2
|
A2-B3-C1
|
A2-B3-C1
|
T (℃)
|
416
|
|
|
396
|
Fc (N)
|
350
|
|
|
292
|
Ra (µm)
|
0.8437
|
|
|
0.5837
|
GRG
|
1.872
|
|
2.3971
|
2.396
|
The improvement in GRG = 0.524
|
The percentage improvement in GRG = 21.87%
|
When the results are evaluated, it is seen that the estimated results are better. It was found that there is a good correlation between the predicted GRG and the experimental results. The improvement in GRG from the initial parameters to the optimum parameters was 0.524 so 21.87%. The values obtained from the validation test showed that the GRG values were compatible with the confidence interval limits. As a result, Taguchi based grey relational analysis methodology for Fc, T and Ra has been successfully applied.