Multi-Objective Optimization of Milling Ti-6Al-4V Alloy for Improved Surface Integrity and Sustainability Performance

Abstract

Ti-6Al-4V is a titanium-based alloy that is widely used in a diverse range of applications, especially in industries such as biomedical and aerospace. Several lubricooling techniques have been introduced to enhance the machinability of these materials. Among them, environmentally friendly strategies are gaining in importance, with sustainability trends rising in manufacturing. The present research investigates the effect of two eco-friendly lubricooling techniques (minimum quantity lubrication and cryogenic cooling), along with other cutting parameters (cutting speed and feed per tooth), on the surface roughness and microhardness of the machined surfaces, which are identified as one of the most frequently implemented indicators of surface integrity in the ball-end milling of the Ti-6Al-4V alloy. In addition, the total electrical energy consumption of the machine tools under different cooling/lubrication conditions was also analyzed. The results obtained showed that cryogenic cooling enhanced milling performance as compared to MQL. Moreover, a multi-objective parameter optimization model integrating the machining responses (surface roughness, microhardness, energy consumption, and productivity) and sustainability metrics (environmental impact, operator’s health and safety, and waste management) was introduced. It was found that cryogenic cooling outperformed the MQL method in terms of both machining performance and environmental impact. An analysis of variance (ANOVA) was carried out to evaluate the significance of each process parameter on the multiple performance index. The results indicate that feed per tooth, cooling method, and cutting speed were significant, with respective contributions of 39.4%, 36.8%, and 22.9%. Finally, the optimal parameter setting was verified through a confirmation test and the results reveal that an improvement was observed in the machining responses and multiple performance index.

1. Introduction

The demand for titanium and nickel-based alloys, with enhanced mechanical properties at elevated temperatures, is constantly increasing in many industries, such as aerospace, medical, chemical, and nuclear, and other engineering areas as well [1]. Ti-6Al-4V is the most widely used titanium-based alloy, accounting for more than half of worldwide production, and 80% of which is utilized in aerospace and medical applications. However, this alloy is generally classified as one of the most difficult-to-cut materials due to its low thermal conductivity and high chemical reactivity [2]. Rapid tool wear, high cutting temperature, and low productivity are major problems while machining Ti-6Al-4V. To overcome these problems and improve the machinability of difficult-to-cut materials, cutting fluids or metal-working fluids are applied. The conventional use of cutting fluids (flood cooling) is widely employed for the machining of difficult-to-cut materials because of their superior cooling/lubrication capabilities [3]. However, mineral-oil-based cutting fluids may cause adverse effects on human health and the environment. Furthermore, the cost of the disposal and treatment of these cutting fluids can be two times higher than their production cost [4]. In addition to workpiece materials and machining parameters, the choice of the lubricooling technique appears to be crucial in terms of the sustainability of the machining process. Hence, the researchers proposed various environmentally friendly and resource-saving cooling/lubrication alternatives, such as dry machining, machining with a minimum quantity of lubricant (MQL), cryogenic machining, and a combination of cryogenic and MQL (hybrid cryoMQL) machining [5].

In the last few years, different eco-friendly cooling/lubrication approaches have been extensively researched to enhance machining performance when milling titanium-based alloys. The experimental investigations by various researchers explored the potential of the MQL method. According to Jamil et al. [6], the application of the MQL method in the high-speed machining of Ti-6Al-4V reduces cutting force, cutting temperature, and residual stresses and improves surface quality and tool life as compared to dry machining. It was demonstrated that a graphene additive improves lubrication performance in the MQL milling of titanium-based alloys [7]. The effects of oil supply rates on milling force, tool wear, and surface quality in the high-speed MQL-assisted milling of Ti-6Al-4V alloy were discussed in [8]. Garcia and Ribeiro [9] utilized a vegetable-oil-based minimal quantity of fluid technique to study the tool life and surface roughness in the end milling of the Ti-6Al-4V alloy. Longer tool life and lower surface quality were reported as compared to dry machining and cutting with a fluid jet. Liu et al. [10] studied the influence of operating MQL parameters, namely nozzle distance and position, air pressure, and quantities of oil consumed, on the cutting force and temperature. Jamil et al. [11] applied a hybrid nanofluid-based MQL technique in the milling of Ti-6Al-4V to find the influence of cutting speed, feed rate, depth of cut, and width of cut on surface quality, energy consumption, tool wear, and chip morphology. Moreover, multi-objective optimization was performed to determine the optimum level of cutting parameters. The effects of various hybrid nanofluid-assisted MQL parameters, including nanofluid concentrations, air pressure, and flow rate, on the surface roughness, cutting temperature, and cutting force in the milling of the difficult-to-cut alloy Ti-6Al-4V were investigated [12]. Bai et al. [13] performed milling experiments on the Ti-6Al-4V alloy using six types of nanofluid. In this study, cutting force, surface quality, and workpiece surface morphology were selected as machinability indicators. The effects of MQL with different concentrations of hexagonal boron nitride nanoparticles on the cutting force and surface roughness of slot-milled titanium alloy were investigated [14].

Cryogenic cooling is also an effective technique that has been widely applied in the milling of titanium alloys. Shokrani et al. [15] compared the impact of cryogenic, dry, and flood cooling in the milling of the Ti-6Al-4V alloy. Regarding tool life, surface quality, power consumption, and specific machining energy, the authors concluded that cryogenic cooling is the most favorable machining environment. Cryogenic cooling leads to significantly improved surface integrity [16] and energy consumption [17] in the end milling of Ti-6Al-4V as compared with dry and conventional cooling. Sadik and Isakson [18] found that cryogenic cooling can significantly improve tool life in the face milling of Ti-6Al-4V as compared to conventional cooling. Lee et al. [19] presented the influence of dry, cryogenic, and cryogenic plus preheated cutting environments on the tool wear, machining forces, and chip morphology when milling Ti-6Al-4V. It is demonstrated that cryogenic treatment offers improved machining performance in terms of longer tool life and reduced overall cost when compared to dry and conventional cutting. Masood et al. [20] proved that the application of cryogenic cooling during the milling of the Ti-6Al-4V alloy elongates the cutting tool life and reduces machining cost in comparison to those obtained under dry machining. Safari et al. [21] analyzed the effect of cutting speed and feed rate in the high-speed end milling of a titanium-based alloy under a cryogenic environment. Albertelli and Monno [22] performed milling tests on Ti-6Al-4V and discovered that cryogenic cooling is at least 25% less energy-demanding than dry and conventional cooling. The face milling of alloy Ti-6Al-4V has been conducted with cryogenically treated and plain WC cutting inserts [23]. Research results revealed that the use of cryogenically treated tools contributed to enhanced surface roughness and reduced cutting force, tool wear, and tool vibration. Li et al. [24] constituted a specific cutting energy model during the cryogenic milling of the Ti-6Al-4V alloy according to the chip formation mechanism.

Nowadays, the performance comparison of environmentally friendly cooling conditions in the milling of titanium and its alloys is a popular topic of interest among researchers. It was reported that cryogenic cooling with MQL improves the machinability characteristics of the Ti-6Al-4V alloy in terms of tool wear, cutting forces, and surface quality as compared to dry, conventional, and MQL conditions [25]. In order to evaluate cutting force and tool wear, Park et al. [26] conducted an experimental investigation on the milling of Ti-6Al-4V alloy under cryogenic, MQL, dry, and wet cooling environments. Jamil et al. [27] compared the performance of the dry, MQL, and cryogenic (LN₂ and CO₂) milling of the Ti-6Al-4V alloy regarding machining characteristics (tool wear, surface roughness, cutting temperature) and sustainability measures (processing time, specific cutting energy, energy consumption, carbon emissions). The results demonstrated that CO₂ outperformed MQL, LN₂, and dry cutting conditions in terms of overall performance. Cai et al. [28] analyzed the potential of dry, supercritical CO₂, supercritical CO₂-based MQL with water-based cutting fluid, and supercritical CO₂-based MQL with oil-on-water droplets cutting fluid during the finish milling of Ti-6Al-4V. The cutting force, cutting temperature, milling stability, and surface roughness morphology were considered as machining evaluators. An et al. [29] investigated tool wear and surface quality in the side milling of Ti-6Al-4V under dry, supercritical CO₂, supercritical CO₂ with antifreeze-water-based MQL, and supercritical CO₂ with oil-on-water-based MQL. MQL, cryogenic, cryoMQL, and laser-assisted machining provide lower tool wear and cutting force than dry and flood cooling in the face milling of the Ti-6Al-4V alloy [30]. Bagherzadeh et al. [31] applied various cooling/lubricating methods, namely, MQL, cryogenic (LN₂ and CO₂), and hybrid (LN₂ + MQL and CO₂ + MQL), during the slot milling process of Ti-6Al-4V. It was found that, in general, LN₂ or LN₂ + MQL outperformed the other methods in terms of tool wear, cutting force, temperature, and surface roughness. The effects of MQL and cryogenic cooling were also analyzed with respect to specific energy consumption [32]. In the experimentation, it was found that the difference between the studied techniques was very small. Park et al. [33] conducted experiments under cryogenic, MQL, LN₂ + MQL, and conventional cooling in the milling of the Ti-6Al-4V alloy. They reported that LN₂ + MQL outperformed the conventional cooling with respect to tool life and cutting force. Wu et al. [34] conducted comparable tests of flood cooling, MQL, and cryogenic MQL milling processes of the titanium alloy Ti-6Al-4V. The results showed that the cryogenic MQL method has a high potential to reduce cutting force and tool wear and to enhance machined surface quality. Ha et al. [35] revealed that the application of the integrated injection of CO₂ and MQL mist in the milling of the Ti-6Al-4V titanium alloy effectively reduces the cutting temperature and tool wear as compared to other lubricooling techniques, namely, dry, cryogenic, and MQL. They also noticed a reduced cost and increased productivity.

The literature review shows that many researchers have studied various cooling/lubricating methods to improve the machinability of titanium alloys. However, the integration of environmental indices of these lubricooling conditions with the common machining process measures is required to understand process performance. Furthermore, multi-objective parameter optimization considering both machinability and sustainability metrics needs to be studied. It can also be seen that limited studies were carried out for the ball-end milling operation. The goal of this paper is to fill the gap in this field. The effects of two lubricooling methods, namely MQL and cryogenic (CO₂), on machined surface integrity (surface roughness and microhardness) and total electrical energy consumption were studied in the ball-end milling of the Ti-6Al-4V alloy at different cutting speeds and feed rates. In addition, a multi-objective cutting parameter optimization model was introduced, which takes surface roughness, microhardness, energy consumption, productivity, carbon emission, operator health and safety impacts, and waste management as optimization objectives. Figure 1 shows the methodology of the present research work.

This paper is organized as follows. In Section 2, the experimental details are described. The effects of cutting conditions, using both cooling/lubrication methods, on machining performance in terms of surface roughness, microhardness, and energy consumption have been evaluated and compared in Section 3. The multi-objective optimization model is presented in Section 4. Finally, Section 5 deals with the conclusions of this study.

2. Experimental Work

2.1. CNC Machine Tool, Workpiece Material, and Cutting Tool

For this investigation, ball-end milling trials were performed on a CNC machining center with a maximum spindle speed of 20,000 rpm and a maximum motor power of 22 kW (NX 6500 II Doosan (DN Solutions, Seoul, Republic of Korea) make with a Fanuc control).

The workpiece material for the research was Ti-6Al-4V alloy with a nominal chemical composition (wt. %) of Al: 5.5–6.75, V: 3.5–4.5, Fe: <0.25, O: <0.2, C: <0.08, N: <0.05, H: <0.01, Ti: remaining. The material has 30–45 HRC hardness with a yield strength of 725–930 N/mm². The workpieces of 62 mm in length, 41 mm in width, and 12 mm in height were pre-machined and utilized for the ball-end milling in the MQL and cryogenic cutting environment.

The cutting tool used in all experiments was a two-flute carbide ball-nosed end mill with a diameter of 10 mm, a flute length of 19 mm, a helix angle of 30°, and a rake angle of 10°. The relative inclination angle between the ball cutter and workpiece significantly influences surface quality. With the purpose of minimizing the plastic deformation of the machined layer due to the tool axis’s zero effective cutting speed, all experimental trials were conducted using a tool inclination angle of 20°. For the case of ball-end milling, the effective cutting speed is computed on the basis of the effective cutter diameter. In the present study, the effective tool diameter was 8.76 mm.

2.2. Cooling Conditions and Parameter Selection

An experimental study was conducted using two eco-friendly cooling/lubrication techniques, namely, MQL and cryogenic cooling. MQL is one of the most popular sustainable techniques that facilitates near-dry machining. In this method, a mist made by droplets of oil and pressurized air is delivered directly into the cutting zone. MQL generally reduces cutting fluid consumption and, consequently, minimizes the cooling/lubrication cost and wastage disposal problems associated with them. This technique has a low environmental footprint and decreases worker health issues. Additionally, the chips produced by this technique are nearly dry and considerably cleaner as compared to flood cooling. On the other hand, MQL has poor cooling capability, chip removal and chip breaking are unsatisfactory, particle emission exists, and MQL spray may cause inhalation problems. In the present research, an MQL system that consists of a source of compressed air, a cutting fluid tank, a flow control system, tubes, and a spray nozzle was utilized. For the MQL system parameters, namely an air pressure of 0.6 MPa and a flow rate of 60 mL/h, were kept constant over all the experimental trials. The nozzle has a diameter of 1 mm and it was fixed about 30 mm away from the tool–chip interface at a position angle of approximately 60° to the tool axis.

Cryogenic cooling is also a popular sustainable lubricooling technique that uses cryogenic gases such as nitrogen and carbon dioxide as a cooling medium. These coolants absorb the heat generated during machining and evaporate quickly into the air without any pollution. The application of this technique favorably affects surface quality and tool life, thereby leading to higher productivity at low energy consumption. Because the chips produced by cryogenic cooling have no oil/fluid/water residue, they can be recycled as scrap metal. This technique is also environmentally safe because the coolant evaporates harmlessly into the air without producing any waste. High capital, operational, and maintenance costs are the main disadvantages of cryogenic cooling. A typical cryogenic cooling system composed of a cylinder of CO₂, a pressure system, a flow control system, tubing, and a nozzle was utilized in the present research. For ball-end milling with cryogenic cooling, a 0.8 mm diameter nozzle was used to supply 12 kg/h CO₂ at 5.7 MPa delivery pressure. The fixed distance between the nozzle and the tool–chip interface was set at 2 mm for all the trials.

In the current work, the Taguchi mixed-level orthogonal array (OA) L₃₆ (2¹ 3²) was used for the experiment design, which implied carrying out 36 trials based on one factor of two levels and three factors of three levels, as listed in

Table 1. Experimental L₃₆ (2¹ 3²) orthogonal design array and results.

No.	Control Factor Level			Machining Outputs				Sustainable Indicators			MPI
	LCT	fz (mm/z)	vc (m/min)	Sa (μm)	HV0.1	E (Wh)	MRR (mm3/min)	CO2 Emission (kgCO2/kWh)	OHS	WM
1.	1	1	1	1.57	375	393.3	211.0	0.1498	2	2	0.4178
2.	1	2	2	1.56	367	237.9	387.6	0.0906	2	2	0.5976
3.	1	3	3	1.54	369	200.2	615.8	0.0763	2	2	0.7462
4.	1	1	1	1.47	380	404.6	211.0	0.1542	2	2	0.4237
5.	1	2	2	1.70	381	241.9	387.6	0.0922	2	2	0.5852
6.	1	3	3	1.70	380	181.9	615.8	0.0693	2	2	0.7827
7.	1	1	1	1.60	373	394	211.0	0.1501	2	2	0.4139
8.	1	2	2	1.68	368	240	387.6	0.0914	2	2	0.5831
9.	1	3	3	1.59	376	181.4	615.8	0.0691	2	2	0.7937
10.	1	1	1	1.59	377	398.7	211.0	0.1519	2	2	0.4142
11.	1	2	2	1.77	392	243.4	387.6	0.0927	2	2	0.5819
12.	1	3	3	1.65	371	179.3	615.8	0.0683	2	2	0.7913
13.	1	1	2	1.73	368	290	301.4	0.1105	2	2	0.4950
14.	1	2	3	1.54	376	199.6	503.9	0.0760	2	2	0.7131
15.	1	3	1	1.58	395	266.2	331.6	0.1014	2	2	0.5534
16.	1	1	2	1.55	365	285.4	301.4	0.1087	2	2	0.5137
17.	1	2	3	1.59	380	196.5	503.9	0.0749	2	2	0.7169
18.	1	3	1	1.51	392	262.5	331.6	0.1000	2	2	0.5634
19.	2	1	2	1.51	399	285.6	301.4	0.1088	1	2	0.6580
20.	2	2	3	1.51	405	202.5	503.9	0.0772	1	2	0.8787
21.	2	3	1	1.45	409	262.8	331.6	0.1001	1	2	0.7103
22.	2	1	2	1.44	403	280.8	301.4	0.1070	1	2	0.6736
23.	2	2	3	1.40	400	198.7	503.9	0.0757	1	2	0.9022
24.	2	3	1	1.40	414	265.9	331.6	0.1013	1	2	0.7160
25.	2	1	3	1.41	422	229.5	391.9	0.0874	1	2	0.7995
26.	2	2	1	1.25	404	317.2	271.3	0.1209	1	2	0.6563
27.	2	3	2	1.29	406	209.3	473.7	0.0797	1	2	0.8860
28.	2	1	3	1.33	417	233.4	391.9	0.0889	1	2	0.8012
29.	2	2	1	1.35	413	319.3	271.3	0.1217	1	2	0.6420
30.	2	3	2	1.46	401	206	473.7	0.0785	1	2	0.8623
31.	2	1	3	1.24	411	241.9	391.9	0.0922	1	2	0.8004
32.	2	2	1	1.16	406	332.6	271.3	0.1267	1	2	0.6604
33.	2	3	2	1.33	419	213.5	473.7	0.0813	1	2	0.8762
34.	2	1	3	1.35	410	238.4	391.9	0.0908	1	2	0.7858
35.	2	2	1	1.18	419	330	271.3	0.1257	1	2	0.6653
36.	2	3	2	1.28	413	215.6	473.7	0.0821	1	2	0.8770

. The Taguchi method is a robust experimental design technique that has been widely used in engineering analysis. The greatest advantages of this method are the saving of effort and time in conducting experiments, reducing the cost, and discovering significant factors quickly. The selected process parameters that are outlined as the controllable factors used in this study are eco-friendly lubricooling techniques (LCT), cutting speed (v_c), and feed per tooth (f_z). The experiments were conducted under two different lubricooling environments (MQL and cryogenic), at three different cutting speeds (70, 90, and 110 m/min), and with three different feed per tooth values (0.07, 0.10, and 0.13 mm/z). In the experimental layout for the L₃₆ OA, numerical values, i.e., 1, 2, and 3, were assigned to a specific level. The other cutting parameters, such as the depth of cut and step-over, were maintained constant at 1 mm and 0.59 mm, respectively. The cutting parameters were selected according to recommendations from the tooling manufacturer.

2.3. Response Measurements

In the present work, the surface quality characteristics of the machined surface were obtained by measuring surface roughness and microhardness. After each experimental trial, the machined surface roughness was measured using the optical, non-contact, and three-dimensional microscope InfiniteFocus SL (Alicona, Graz, Austria). Traditionally, the measurement of surface topography using tactile stylus-based profilometers resulted in a two-dimensional profile. Some of the technical shortcomings of this method are that the measurement position has a serious influence on the obtained results, the sample surface may be deformed or damaged due to the contact principle of measurement, the measurement process is time-intensive, on-line measurement is difficult, etc. The aforementioned drawbacks are especially noticeable when the measurement of micro/nano-scale manufactured parts is required. The InfiniteFocus SL measuring device enables complete 3D surface analysis and it is not subjected to those limitations; therefore, it appears to be an adequate alternative for the measurement of surface quality. In the present research, the average height of the selected area (S_a) was selected to evaluate surface quality. The vertical resolution was 50 nm and the lateral resolution was 3 µm. The roughness parameters used were measured in the direction perpendicular to the feed.

Microhardness measurements on machined samples were performed by the DuraScan G5 (Emco, Salzburg, Austria) digital hardness tester. The test load of 100 g was applied. Each measurement was repeated five times at different locations of the milled surface, and the average value of each output was reviewed as a microhardness criterion.

The total electrical energy consumption of the CNC machine tool was monitored and measured through a three-phase energy and power disturbance analyzer MAVOWATT (Gossen Metrawatt, Nurnberg, Germany).

3. Results and Discussion

The results of the experimentation using L₃₆ Taguchi OA are shown in Table 1. Firstly, the mathematical models of surface roughness, microhardness, and total electrical energy consumption were developed. Then, an analysis of variance (ANOVA) was used to evaluate the significance of the regression model as well as individual model coefficients. Finally, 3D response surface plots were constructed to investigate the influences of lubricooling techniques and milling parameters on the responses.

3.1. Results and Analysis of Surface Roughness

As mentioned earlier, the significance of the regression model and the influence of each machining factor on the individual response is determined by using ANOVA. The linear terms, interaction terms, and square terms are listed as the source of contribution in an ANOVA table. The ANOVA operates based on parameters, namely the sum of squares, mean square, F-value, and p-value. The sum of squares denotes the deviation from the mean, which is derived from that source, while mean square is the sum of squares divided by its associated degrees of freedom. F-value is a test for comparing the model variance with the residual (error) variance. If the variances are close to the same, the ratio will be close to one and it is less likely that any of the factors have a significant effect on the response. F-value is calculated using the model mean square divided by the residual mean square. Finally, the p-value represents the statistical significance to a confidence interval of 95%.

An ANOVA for an interval of confidence of 95% was applied to quantify the effect of each process parameter and their interaction on the surface roughness (

Table 2. Results of the ANOVA table for the average height of the selected area (S_a).

Source	Sum of Squares	df	Mean Square	F-Value	p-Value	Significance
Model	0.6495	3	0.2165	28.86	<0.0001	Significant
LCT	0.5894	1	0.5894	78.58	<0.0001	Significant
fz	0.0224	1	0.0224	2.99	0.0935	Not significant
fz2	0.0377	1	0.0377	5.02	0.0321	Significant
Residual	0.24	32	0.0075
Total	0.8895	35

). The model can be interpreted according to the F-value and p-value. The model F-value of 28.86 implies the model is significant. There is only a 0.01% chance that an F-value this large could occur due to noise. The p-values for the model variables assist in determining the level of significance (<0.05) of individual, squared, and interaction effects of each factor. The Fisher’s F-test was also used to determine the significance of the factors. If the F-value is higher, the significance is higher for the model term. It is worth emphasizing that the lubricooling technique revealed the most significant effect on the average height of the selected area. Feed per tooth has much less contribution, whereas cutting speed does not have a statistical significance.

The reduced quadratic mathematical model of the average height of the selected area (S_a) for MQL and the cryogenic technique is defined in Equation (1) and Equation (2), respectively.

$$S_{aMQL}=0.78879+16.2645f_z-76.2298{f_z}^2,$$

(1)

$$S_{aCryo}=0.53288+16.2645f_z-76.2298{f_z}^2,$$

(2)

The 3D surface graphs for the average height of the selected area as a function of cutting speed and feed per tooth for the two eco-friendly lubricooling techniques are shown in Figure 2. The results exhibited a considerable decrease in surface roughness with cryogenic cooling. Similar results in the machining of titanium alloys were reported by other authors [27,31]. This can be explained by the improved penetration of coolant into the tool–chip interface and by the reduced formation of built-up edge on the cutting tool. In addition, for both lubricooling techniques, surface roughness values slightly decreased with decreasing feed per tooth. This fact is confirmed by the metal cutting theory because increasing feed per tooth results in greater heat generation, which, in addition to the low thermal conductivity of Ti-6Al-4V, caused heat elevation at the cutting zone, and then decreased surface finish is generated. The three-dimensional (3D) topographies under both lubricooling environments are shown in Figure 3. The minimum average height of the selected area under MQL and cryogenic cooling was achieved for experiment No. 4 and No. 32, respectively. The results obtained show that, using both lubricooling techniques, the typical surface ball-end milling pattern was generated. A similar phenomenon was also observed under other machining parameters. However, as shown in Figure 3, it was observed that the surface texture has more obvious peaks and valleys with a larger average height of the selected area (S_a = 1.47 μm) in MQL machining when compared to that of the texture with a lower average height of the selected area (S_a = 1.16 μm) obtained by cryogenic cooling.

3.2. Results and Analysis of Microhardness

The statistical analysis for the microhardness is given in

Table 3. Results of the ANOVA table for the microhardness (HV_0.1).

Source	Sum of Squares	df	Mean Square	F-Value	p-Value	Significance
Model	10,142.1	4	2535.51	50	<0.0001	Significant
LCT	5468.45	1	5468.45	107.38	<0.0001	Significant
vc	84.38	1	84.38	1.66	0.2066	Not significant
fz	66.67	1	66.67	1.31	0.2603	Not significant
vc × fz	452.23	1	452.23	8.92	0.0055	Significant
Residual	1572.17	31	50.72	1.99

. The model F-value of 50 implies the model is significant. There is only a 0.01% chance that an F-value this large could occur due to noise. The p-value of the fitted model is less than 0.0001, which indicates that the obtained model is statistically significant. In terms of p-value and F-value, the lubricooling technique is the most dominant control factor for the microhardness. It has been observed that the main effects of cutting speed and feed per tooth are not significant, whereas the interaction effect between them is significant.

The regression equation for the microhardness in terms of lubricooling technique, cutting speed, and feed per tooth is shown in Equations (3) and (4) below.

$$H{V}_{MQL}=284.025+1.17v_c+865.28f_z-10.231v_c{f}_z,$$

(3)

$$H{V}_{Cryo}=312.488+1.17v_c+865.28f_z-10.231v_c{f}_z,$$

(4)

Figure 4 shows the 3D plot representing microhardness evolution as a function of the combination of process parameters under different sustainable lubricooling environments. From this result, it can be concluded that cryogenic cooling induces higher values of microhardness below the machined surface compared to MQL machining. Similar results have been reported by Rotella et al. [36]. In the ball-end milling of difficult-to-cut materials, such as the Ti-6Al-4V titanium-based alloy, a significant amount of cutting heat is generated, resulting in changes in the mechanical properties beneath the machined surface, such as variations in the hardness values. The cryogenic environment, compared with MQL, decreases the cutting temperature remarkably and, subsequently, leads to strain hardening on the surface and subsurface of machined parts. The results also show a slightly increased microhardness with higher values of cutting speed, independently of the lubricooling technique. Because of the increased temperature of the shear zone, this phenomenon causes a phase transformation and, as a result, rapid cooling, which enhances the microhardness.

3.3. Results and Analysis of Electrical Energy Consumption

Table 4. Results of the ANOVA table for the electrical energy consumption (E).

Source	Sum of Squares	df	Mean Square	F-Value	p-Value	Significance
Model	1.45∙105	5	28,995.6	715.02	<0.0001	Significant
vc	44,290	1	44,290	1092.18	<0.0001	Significant
fz	89,279.6	1	89,279.6	2201.6	<0.0001	Significant
vc × fz	6922.2	1	6 922.2	170.7	<0.0001	Significant
vc2	561.1	1	561.1	13.84	0.0008	Significant
fz2	3925	1	3925	96.79	<0.0001	Significant
Residual	1216.6	30	40.6
Total	1.462∙105	35

represents the ANOVA table of the second-order model proposed for the total electricity energy consumption response. The model F-value of 715.02 implies the model is significant. There is only a 0.01% chance that an F-value this large could occur due to noise. It can be seen that all the effects have a p-value less than 0.05. These significant effects, arranged in order of importance, are the main effect of the feed per tooth, the main effect of the cutting speed, the interaction effect between them, and the quadratic effects of the feed per tooth and cutting speed. It can also be underlined that lubricooling technique does not show a significant effect on the electrical energy consumption the of machine tool.

The quadratic model of total energy consumption in terms of cutting speed and feed per tooth is presented as follows.

$$\begin{array}{cc}\hfill E_{MQL}=E_{Cryo}=& 1364.523-9.383v_c-10075.28f_z+\hfill \\ \hfill & 34.667v_c{f}_z+0.02094{v_c}^2+24611.11{f_z}^2\hfill \end{array},$$

(5)

As shown in Figure 5, concerning cooling/lubricating conditions, an identical trend has been shown with the results of electrical energy consumption. Electric power demand associated with the use of delivery units for both lubricooling techniques was not considered due to the fact that these units are not an integrated part of the machine tool, whereas the energy consumption of the machine tool was measured through a device attached to the main bus of the CNC machine tool electric cabinet. Hence, in this study, the direct electrical energy demand was analyzed. It can be observed that there is no difference between the two lubricooling techniques regarding the machine tool’s electrical energy consumption, and that was confirmed by the results obtained from the ANOVA. Cryogenic machining enhances heat dissipation from the cutting zone and, therefore, reduces cutting forces and, indirectly, the overall energy consumption required to produce the machined components. On the other hand, the application of MQL improves energy consumption due to reduced friction at the cutting tool–workpiece material interface. Accordingly, both lubricooling techniques can be classified as eco-friendly cooling/lubrication approaches that save energy resources while maintaining high machining efficiency. According to Figure 5, it can be observed that an increase in feed per tooth and cutting speed causes a decrease in energy consumption in the ball-end milling of the Ti-6Al-4V alloy. Hence, the minimum value of this response was obtained when the levels of both cutting parameters were at their highest levels. This phenomenon can be explained by the fact that the higher the feed per tooth and cutting speed, the higher the MRR and, consequently, the time needed for the machining operation is reduced [17].

4. Sustainability-Based Optimization Model

One of the main goals of the present research is to establish an optimization model that integrates the machining performances with a sustainability assessment with the purpose of searching for improvement options using lubricooling techniques. The sustainability assessment model proposed by Hegab et al. [37] was utilized. The key steps of this method are as follows:

(i) The determination of the machining responses and the calculation of the sustainable factor:

$$S{F}_{kpj}=\left\{\begin{array}{c}{I}_{kp}\cdot M{r}_j, if I_{kp} and M{r}_j are both LB or HB\\ {\displaystyle \frac{Min\left(I_{kp},M{r}_j\right)}{Max\left(I_{kp},M{r}_j\right)}}, otherwise\end{array}\right.,$$

(6)

where SF_kpj is the sustainable factor for each machining test (n), I_kp is the value for each sustainable indicator (p), and M_rj is the value for each machining response (j).

(ii) The normalization and calculation of the sustainable index:

$$S{I}_{kpjn}=\left\{\begin{array}{c}{\displaystyle \frac{Min\left(S{F}_{kpj}\right)}{S{F}_{kpjn}}}, if S{F}_{kpj} is based on LB\\ {\displaystyle \frac{S{F}_{kpjn}}{Min\left(S{F}_{kpj}\right)}}, if S{F}_{kpj} is based on HB\end{array}\right.,$$

(7)

where SI_kpjn is the sustainable index for each machining test.

(iii) The calculation of the weighted sustainability index:

$$WS{I}_{kpjn}=SM{W}_k\cdot MQ{W}_j\cdot I{W}_{kp}\cdot S{I}_{kpjn},$$

(8)

where WSI_kpjn is the weighted sustainable index for each cutting test, SMW_k is the weighting importance factor for each sustainable metric (k), MQW_j is the weighting importance factor for each machining response (j), and IW_kp is the weighting importance factor for each sustainable indicator (p).

(iv) The calculation of the total weighted multiple performance index:

$$MP{I}_n={\displaystyle \sum _{k=1}^{NM}{\displaystyle \sum _{p=1}^{N_k}{\displaystyle \sum _{j=1}^{M}WS{I}_{kpj}}}}$$

(9)

where MPI_n is the total weighted sustainable index for each cutting test, NM is the number of studied sustainable metrics, N_k is the number of selected indicators for each sustainable metric, and M is the number of selected machining responses.

In this work, the objective of optimization is to minimize the environmental footprint associated with energy consumption without deteriorating the surface roughness of machined parts, microhardness, and material removal rate, as these might lower productivity. In addition to previously studied machining outputs (i.e., the average height of the selected area, microhardness, and energy consumption), the sustainability metrics were also considered to evaluate each lubricooling approach. The proposed sustainability indicators for this study fall into three categories, namely, environmental impact (represented by CO₂ emissions), operator’s health and safety (OHS), and waste management (WM). The considered indicators would be capable of quantifying sustainability based on the results acquired in the previous sections.

CO₂ emissions were determined based on the electrical energy consumed in each machining trial using a standard established emission intensity, as per Narita et al. [38], whereas the operator’s health and safety and waste management were added as a qualitative metric, which are lower than the better indicators. For the MQL and cryogenic technique, the operator’s health and safety factors were assigned as 2 and 1, respectively. Despite the fact that, in MQL machining, a small quantity of oil is used, from a health point of view, air–oil mist creation is one of the major drawbacks of this method and cannot be ignored, in particular, if applied in industry environment conditions where the machining processes are likely to continue for several hours. In consequence, the operator’s breathing problems were observed. On the other hand, cryogenic machining is identified as a safer option due to the complete elimination of oil-based cutting fluids. In terms of the waste management of used coolants and chip generation, equal values were assigned for both cutting environments. MQL and cryogenic techniques do not leave the residue of cutting fluids that need to be disposed of either. Moreover, chips produced under MQL and cryogenic cooling are practically clean and easy to recycle. In the present study, equal weights were assigned to each metric, i.e., machining outputs and sustainable indicators. The design for the sustainability assessment as well as the assessment results for all experimental trials are provided in Table 1. The highest multiple performance index (MPI = 0.9022) was observed in experimental trial No. 23 using cryogenic cooling at a cutting speed of 90 m/min and a feed per tooth of 0.13 mm/z. For the MQL cutting environment, the highest MPI of 0.7937 was noticed in machining test No. 9 at a cutting speed of 110 m/min and a feed per tooth of 0.13 mm/z.

In addition to a sustainability assessment, a Taguchi-based optimization of process parameters is also presented in this section. This method is a well-known and powerful tool that provides a simple, efficient, and systematic approach for the optimization of experimental designs. An additional advantage is the ease of intermixing numerical and categorical variables in optimization problems. However, the traditional Taguchi approach is primarily designed to improve process optimization with a single performance characteristic, i.e., single objective optimization. Hence, the analyzed multiple objectives must be converted into an equivalent single response. In this context, MPI can serve as the single objective function for optimization, with the goal of maximizing it. The S/N ratios were calculated using the condition “the higher the better” for each of the experimental trials from the L₃₆ orthogonal array. The average values of S/N ratios for each level of the process parameters are shown in Figure 6. Based on the obtained results, cryogenic cooling and cutting speed as well as feed per tooth at the highest level indicate the optimum levels of process factors for maximum MPI.

ANOVA was also performed to evaluate the relative significance of the machining parameters on the multiple performance index (

Table 5. Results of the ANOVA table for the multiple performance index (MPI).

Source	Sum of Squares	df	Mean Square	F-Value	p-Value	Significance
LCT	52.07	1	52.07	1366.62	<0.0001	Significant
vc	32.39	1	32.39	850.16	<0.0001	Significant
fz	55.67	1	55.67	1461.02	<0.0001	Significant
Residual	1.22	32	0.038
Total	141.35	35

). According to the F-value of each individual, the level of the significant parameter was in the order of feed per tooth, lubricooling technique, and cutting speed. Moreover, the percentage contribution of the main effects has also been calculated from the sum of squares of each factor. According to the obtained results, feed per tooth was the major factor affecting the MPI with a 39.4% contribution, followed by lubricooling technique and cutting speed with 36.8% and 22.9% contributions, respectively. The remaining (0.9%) effects were caused by noise or uncontrollable factors.

After determining the optimal settings of machining parameters, a confirmation test was conducted to evaluate the accuracy of the proposed approach and to validate the experimental results. The confirmation test results carried out at the initial (Cryo, v_c = 90 m/min, f_z = 0.13 mm/z) and optimum factor levels (Cryo, v_c = 110 m/min, f_z = 0.13 mm/z) derived by the considered method are listed in

Table 6. Results of the confirmation test.

Outputs	Initial Parameter Setting	Optimal Parameter Setting	Improvement (%)
Sa (µm)	1.40	1.39	0.7
HV0.1	400	412	3
E (Wh)	198.7	184.3	7.2
MRR (mm3/min)	503.9	615.8	22.2
MPI	0.9022	0.9643	6.9

. As illustrated, a significant improvement in MRR by 22.2%, improvement in energy consumption by 7.2%, an increase in the microhardness by 3%, and a minor reduction in the average height of the selected area by 0.7% were found. Moreover, it is shown clearly that the improvement of MPI from initial to optimal cutting conditions was 6.9%.

5. Discussion

The amount of cutting fluid used in machining operations has been diminishing recently as a result of environmental sustainability requirements. Recent trends indicate that substantial efforts are being made to investigate eco-friendly lubricooling techniques on process performance, particularly in the machining of hard-to-cut materials. One of the most important challenges when machining difficult-to-cut materials is to improve the surface integrity of the machined parts, which is essential for their functional performance. MQL and cryogenic machining have been proposed to resolve the cutting fluid problem and to improve productivity as well as product quality. Thus, the present paper follows the same pattern, and the machining performance investigations, including the sustainability assessment, were carried out under MQL and cryogenic cooling conditions. The experimental results showed that the cryogenic cooling has a positive effect on the surface integrity characteristics of machined components such as surface roughness and microhardness. Exposing the workpiece material and cutting tool to very low temperatures in cryogenic machining has an impact on its mechanical properties. The results of the electrical energy consumption of the machine tool reveal that this performance was not influenced by the lubricooling technique.

To obtain the optimal and sustainable machining conditions for the milling of the Ti-6Al-4V alloy, it is essential to consider the technological responses alongside sustainability metrics. In other words, there is a need to develop a multi-objective optimization method based on the trade-offs between quality, productivity, and sustainability. Considering that machine tools are the leading sources of energy consumption in the industrial sector, reducing carbon emissions is an obligatory component of sustainable manufacturing. Numerous studies have confirmed the relationship between various sustainable metrics and cutting parameters, with a common conclusion that the optimum choice of cutting parameters can considerably reduce the environmental footprint of machine tools. The carbon emissions are directly related to the energy consumed during the machining process, and energy consumption is highly influenced by MRR. Higher values of cutting parameters are preferable to minimize machining time and, consequently, reduce energy consumption and carbon emissions as well. Cryogenic cooling has proven to be have promising potential for employing higher cutting speeds and feed rates even in machining various kinds of difficult-to-cut materials. When the environmental impacts of MQL and cryogenic machining are compared, it can be stated that cryogenic cooling is more sustainable due to the complete elimination of oil-based cutting fluids.

6. Conclusions

This study investigates the machinability of titanium-based alloy Ti-6Al-4V in the ball-end milling process under two eco-friendly lubricooling approaches, namely, MQL and cryogenic cooling. Moreover, multi-objective parameter optimization considering machinability and sustainability metrics was conducted. The following are the conclusions obtained from the present study:

• The lowest values of the average height of the selected area were obtained under cryogenic cooling owing to the enhanced penetration of the coolant into the tool–chip interface and, consequently, this reduced the formation of built-up edge. For both cooling/lubricating conditions, lower feed per tooth leads to better surface quality.
• By comparing two lubricooling techniques, it can be observed that cryogenic cooling induces a higher microhardness on the machined surface, as compared to MQL, due to a significant decrease in the cutting temperature, which favors hardness increase during the machining process.
• No noticeable difference was observed between the two lubricooling techniques regarding the machine tool’s electrical energy consumption. This study revealed that higher levels of cutting speed and feed per tooth caused a decrease in energy consumption due to the reduced time needed to complete the machining operation.
• The developed mathematical models of the average height of selected area, microhardness, and electrical energy consumption are adequate according to statistical analysis. The ANOVA for the average height of the selected area and microhardness revealed that these machinability metrics were primarily influenced by the lubricooling technique. It can also be underlined that feed per tooth and cutting speed were major influencing factors on energy consumption.
• A multi-objective optimization model was established to balance the machinability metrics (surface quality, microhardness, energy consumption, and MRR) and sustainability metrics (carbon emissions, operator health and safety impacts, and waste management). The optimal setting of process parameters that maximizes the multiple performance index was found to be as follows: cryogenic cooling, v_c = 110 m/min, f_z = 0.13 mm/z. The main contribution percentages for feed per tooth, eco-friendly cutting environment, and cutting speed to multiple performance index in the ball-end milling of the Ti-6Al-4V alloy were 39.4%, 36.8%, and 22.9%, respectively.
• Confirmatory test reveals an improvement of 6.9% for the multiple performance index from initial to optimal levels of process parameters, which is satisfactory.
• The implementation of the Taguchi method based on orthogonal arrays not only provides an effective tool to reduce the time and cost of experimentation, but also enhances efficiency.