Cutting tool wear
The analysis of cutting tool wear revealed the main types of wear in the tool, flank and crater wear.
Figure 4 shows the progression of flank surface wear at speeds of 60 and 90 m.min− 1 and feed rates of 0.08 and 0.13 mm.rev− 1 when machining Co-Cr-Mo ASTM F75. We can note that the cutting speed and feed rate combination at high values increases the flank wear rate (VBc).
We can infer that the change from 0.08 to 0.13 mm.rev− 1 in the feed rate is more significant in the flank wear rate when compared with changes in the cutting speed from 60 to 90 m.min− 1. Flank wear increases by about 50% when the feed rate changes, compared to increase cutting speed.
Similarly, Fig. 5 shows the crater wear (KT) on the rake surface of the inserts. We can see a tendency towards crater wear as the cutting parameters increase, indicating the predominant of this wear type in ASTM F75 cobalt alloy turning.
A possible explanation for the predominance of crater wear is the combination of high temperature in the cutting region, intense friction between the tool and the chip, loss of tool coating during the process, increased cutting force with varying feed, in addition to the affinity between the chemical elements of the tool and the workpiece. Baron at al. [18] reported that physicochemical interaction between the cutting tool and the workpiece results in the adhesion of the biomedical Co-Cr-Mo alloy on the flank and rake surfaces. Thus, this adhesion facilitates the removal of hard particles in the tool as the metallic binder dissolves from the core of the cutting tool.
Whereas the flank and crater wear responses are correlated, we performed a multivariate analysis of variance (MANOVA) on these wear patterns and cutting parameters, based on Pearson's correlation with an r-index of 0.746 and a p-value of 0.033 to assess if there is a significant difference. Table 5 summarizes the results of MANOVA. We observed a statistically significant difference in the feed rate parameter for the two types of wear in the cutting tools (p ≤ 0.05).
Table 5
MANOVA statistical analysis for VBc (mm) and KT (mm)
Source
|
Criterion
|
Value
|
F
|
DF1
|
DF2
|
p
|
\(f\) (mm.rev− 1)
|
Wilks Lambda
|
0.055
|
25.51
|
2
|
3
|
0.013
|
|
Lawley-Hotelling
|
17.00
|
25.51
|
2
|
3
|
0.013
|
|
Pillai’s Trace
|
0.944
|
25.51
|
2
|
3
|
0.013
|
\({v}_{c}\) (m.min− 1)
|
Wilks Lambda
|
0.259
|
4.29
|
2
|
3
|
0.132
|
|
Lawley-Hotelling
|
2.860
|
4.29
|
2
|
3
|
0.132
|
|
Pillai’s Trace
|
0.741
|
4.29
|
2
|
3
|
0.132
|
\({v}_{c}\) (m.min− 1) × \(f\) (mm.rev− 1)
|
Wilks Lambda
|
0.530
|
1.33
|
2
|
3
|
0.387
|
Lawley-Hotelling
|
0.884
|
1.33
|
2
|
3
|
0.387
|
Pillai’s Trace
|
0.469
|
1.33
|
2
|
3
|
0.387
|
The influence of the feed rate is well-observed in Fig. 6. It is perceptible that the change in the level from 0.08 to 0.13 mm.rev− 1 is more expressive than cutting speed to flank and crater wear. This predominance may be a result of temperature increases in the cutting region and the stresses generated on the tool rake surface when the feed rate raised. In addition, Cobalt alloys are well-known for low thermal conductivity; hence, increased thermal energy in the cutting region causes a higher wear rate [3, 6, 9, 11].
The changes in the cutting parameters can cause different wear mechanisms, in this way, we performed a morphological characterization on the inserts under all tested conditions to identify the main wear mechanisms in the tools. Figure 7 shows the microscopy and the mass composition values of the elements identified by EDXA. We observe abrasive marks and workpiece material adhered to the tool flank surface
In spectrum 1 of Fig. 7(a.b) is noted evident and well-defined abrasive grooves on flank surface whereas in spectrum 1 of Fig. 7 (c.d) the wear surface is rougher. In both cases, EDXA tests confirmed the presence of adhered material on cutting tool flank. The material adhesion on tool can occur due to the extrusion of workpiece material or chip between the cutting edge and workpiece. Also, the perceptible abrasive grooves on adhered material probably are caused by the continuity of cutting process after the workpiece material adhesion on the worn surface. In this case, the tool hard particles extruded by the workpiece/tool interface, forming the abrasive grooves in the flank wear and adjacent regions, a phenomenon called attrition [24].
As previously seen, the change in cutting speed and feed rate increases the worn area which causes different wear mechanisms. Based on the microscopy shown in Fig. 7, we illustrated the tool's cutting-edge profile in Fig. 8. When we compare only the influence of cutting speed is observed a subtle increase in the crater wear (Fig. 8a.b). On the other hand, when we maintain the cutting speed and increase the feed, this crater wear becomes pronounced (Fig. 8c.d). Besides, with the extreme cutting parameters (Fig. 8d), the crater wear is so deep, which can have weakened the tool, accelerating the flank wear rate.
Combining the low thermal conductivity of the workpiece material and the thermal energy generated may have facilitated the crater wear mechanism. In this way, we can observe that the cutting parameters influenced on crater wear formation, mainly the feed rate.
Figure 9 shows the morphology and mass composition values of the EDXA-identified elements on the rake surface worn.
On spectrum 2, 3 and 4 of Fig. 9(a) and spectrum 1 and 3 of Fig. 9(c), we see the workpiece material adhesion on the worn surface at a cutting speed of 60 m.min− 1. Moreover, in spectrum 1 of Fig. 9(a), we identified the exposed cutting tool substrate and other workpiece elements dispersed in the analyzed region.
While in spectrum 2 of Fig. 9(c) and spectrum 3 of Fig. 9(d), we see material encrusted in the layer adhered to the worn rake surface. According to EDXA of the mentioned points, these particles are rich in Molybdenum, which can be responsible for the abrasion grooves identified in flank and crater wear, as these are precipitates found in the matrix of the Co-Cr-Mo alloy ASTM F75.
This precipitates presence in the crater region is evidence that can help explain the mechanical abrasion grooves on the worn areas of the inserts rake surface. Even with a larger area between the tool and the chip present in the secondary shear zone to dissipate heat, the high temperature in the cutting region makes it easier to remove the tool coating revealing its substrate due to abrasive wear. As the tool core is a cobalt binder, the chemical affinity between the chip and the core tool can cause crater wear to appear. Furthermore, saturation on the adhesion zone favors the adhesion of the workpiece material until the continuous chip formation flow causes surface renewal and constant tool degradation [18].
Tool life
As previously seen, the cutting parameters variation directly impacts the cutting forces and the generated temperature on the cutting edge of the tools. Increasing temperature and forces generated in the turning process significantly influenced the cutting tool's life.
Through MANOVA analysis, we see the feed rate is the cutting parameter that most influences tool wear (Table 5). We plotted in Fig. 10 the tool life versus feed rate for all cutting speeds tested.
The tool life decreases as the feed rate increases at both cutting speeds. Then we can consider a negative correlation between them. We obtained a reduction higher than 50% in tool life when cutting with the feed rate of 0.13 mm.rev− 1. Therefore, we conclude that a higher feed rate is not recommended for turning the Co-Cr-Mo ASTM F75 alloy due to increased wear caused. Tool life and workpiece surface quality are also affected.
Surface roughness
Figure 11 shows average roughness (Sa) of the selected area versus the volume of material removed (Q) for all cutting parameters tested. We observed similar behavior when changing the cutting parameters, the surface roughness value was near-constant until a determined amount of removed material; after that, it decreased. The exception was for cutting speed of 90 m.min− 1 and feed rate of 0.13 mm.rev− 1.
This significant reduction in surface roughness after 3670 mm³ material removal may be related to macroscopic mischaracterization of cutting tool geometry caused by wear. Furthermore, the increase in the feed rate from 0.08 to 0.13 mm.rev− 1 is more significant than the variation in cutting speed since the surface roughness values are equal when considering the standard deviations at the measurement points.
In the case of cut parameter vc of 90 m.min− 1 and f of 0.13 mm.rev− 1, the accentuated wear of the tool can have generated a different behavior in the surface roughness curve. Change in nose radius tool caused by flank wear and increase in effective chip rake angle induced by crater wear interferes significantly in the formation of the machined surface. Thus, when flank and crater wear is combined, the workpiece is cut by a tool withal a highly negative rake face, Fig. 8(d).
To support the hypotheses of surface roughness formation, we submitted the Sa roughness responses of groups 1670, 2890, 3670, and 4030 mm³ removed material at an ANOVA of repeated measures. Thus, we can verify if there is a statistical difference between Sa and the cutting parameters tested, and we note in Table 6a significant interaction (p ≤ 0.05) between them.
Table 6
Repeated Measures ANOVA for Sa (µm) within subjects effects
Source
|
Sum of Squares
|
DF
|
Mean Square
|
F
|
p
|
Roughness Sa
|
0.00751
|
3
|
0.00250
|
0.868
|
0.484
|
Roughness Sa ×\(vc\)
|
0.07558
|
3
|
0.02519
|
8.737
|
0.002
|
Roughness Sa ×\(f\)
|
0.03689
|
3
|
0.01230
|
4.265
|
0.029
|
Roughness Sa × \(vc\) ×\(f\)
|
0.06455
|
3
|
0.02152
|
7.462
|
0.004
|
Residuals
|
0.03460
|
12
|
0.00288
|
-
|
-
|
Considering the interaction between Sa roughness and cutting parameters responses, we applied Tukey's multiple comparison procedure to seclude which groups differ from others. Table 7 depicts only cases with a significant difference between the compared groups.
Table 7
Tukey test for the interaction of Sa × vc × f
Comparisons
|
Mean Difference
|
DF
|
ptukey
|
Roughness
|
vc
|
f
|
|
Roughness
|
vc
|
f
|
1670 mm³
|
60
|
0.08
|
―
|
1670 mm³
|
90
|
0.13
|
-0.4855
|
4
|
0.011
|
1670 mm³
|
60
|
0.13
|
―
|
1670 mm³
|
90
|
0.13
|
-0.3585
|
4
|
0.033
|
1670 mm³
|
90
|
0.13
|
―
|
2890 mm³
|
60
|
0.08
|
0.4525
|
4
|
0.022
|
1670 mm³
|
90
|
0.13
|
―
|
2890 mm³
|
90
|
0.13
|
0.332
|
4
|
0.025
|
1670 mm³
|
90
|
0.13
|
―
|
3670 mm³
|
60
|
0.08
|
0.4155
|
4
|
0.023
|
1670 mm³
|
90
|
0.13
|
―
|
4030 mm³
|
60
|
0.08
|
0.5245
|
4
|
0.009
|
4030 mm³
|
60
|
0.08
|
―
|
4030 mm³
|
90
|
0.13
|
-0.352
|
4
|
0.041
|
Based on qualitative analyses of Fig. 11 and Tukey's test, it is possible to affirm that the change in cutting parameters significantly affects the roughness response only when the tool is sharp. Furthermore, even considering the wear of the cutting tool, the mildest parameters (vc = 60 m.min− 1 and f = 0.08 mm.rev− 1) caused the lowest surface roughness values.
After confirming the difference in surface roughness values for the chip volume generated in the turning process, we performed an ANOVA of repeated measures between the cutting parameters considering all Sa roughness responses of groups 1670, 2890, 3670, and 4030 mm³ of removed material. We could verify if there were differences between the cutting speeds and feed rates tested. ANOVA showed a statistically significant interaction of cut parameters (p ≤ 0.05), Table 8.
Table 8
ANOVA statistical analysis for Sa (µm) between subjects effects
Source
|
Sum of Squares
|
DF
|
Mean Square
|
F
|
p
|
\(vc\) (m.min− 1)
|
0.1856
|
1
|
0.18559
|
46.25
|
0.002
|
\(f\) (mm.rev− 1)
|
0.0965
|
1
|
0.09647
|
24.04
|
0.008
|
\(vc\times f\)
|
0.0371
|
1
|
0.03706
|
9.24
|
0.038
|
Residuals
|
0.0161
|
4
|
0.00401
|
-
|
-
|
According to ANOVA, we apply Tukey's multiple comparisons procedure to isolate which cutting speed and feed rate levels differ, Table 9. Thus, we affirm that only the surface roughness values generated by the smooth parameters (vc = 60 m.min− 1 and f = 0.08 mm.rev− 1) are statistically different, i.e., the combination of the other cutting parameters promotes the same effect on the surface roughness Sa.
Table 9
Tukey test for the interaction of vc × f
Comparisons
|
Mean Difference
|
DF
|
ptukey
|
vc
|
f
|
|
vc
|
f
|
60
|
0.08
|
―
|
60
|
0.13
|
-0.1779
|
4
|
0.017
|
60
|
0.08
|
―
|
90
|
0.08
|
-0.2204
|
4
|
0.008
|
60
|
0.08
|
―
|
90
|
0.13
|
-0.2621
|
4
|
0.004
|
60
|
0.13
|
―
|
90
|
0.08
|
-0.0425
|
4
|
0.588
|
60
|
0.13
|
―
|
90
|
0.13
|
-0.0842
|
4
|
0.171
|
90
|
0.08
|
―
|
90
|
0.13
|
-0.0418
|
4
|
0.599
|
When we plot the interaction of factors tested at the four-measurement points roughness Sa (Fig. 12), we see that the surface roughness depends on the influence of cutting speed and feed rate parameters. Furthermore, at the beginning and end of the turning process, there is no interaction of cutting parameters.
Knowing there is no interaction of cutting parameters at the beginning and end of tool life. We performed a morphological analysis of the turned surfaces to relate the roughness Sa with tool wear. In Fig. 13, we observed the surface roughness generated by opposites conditions of the cutting process for the different volumes of chips removed.
Figure 13a shows that milder cutting parameters generated a gradual increase between valleys and peaks of the formed surface until 3670 mm³ of the removed chip, followed by a decrease in the Sa value at 4030 mm³. These lower roughness surface values can have occurred due to the change in the insert corner radius given by flank wear.
On the other hand, when using severe cutting parameters (Fig. 13b), we observe a change in feed marks on the surface and a decrease in height between valleys and peaks during the cutting process. This phenomenon occurs because the tool wear rate at a given point favors the reduction of roughness values. However, at 4030 mm³, insert wear increases abruptly, causing an increase in the height of the peaks due to the tool macro geometry re-characterization.
Figure 14 shows a detailed analysis of the turned surfaces with vc of 90 m.min− 1 and f of 0.13 mm.rev− 1. We can notice the presence of embedded particles on the workpiece surface. EDXA studies in the particles evince agglomerates rich in molybdenum, probably M23C6 carbides that were not carried away by the cutting action.
Morphological analyses allow us to conclude that the surface roughness value increases as a function of the generated surface's deformation level, which is influenced by the wear suffered by the cutting tool.