The Effect of MoS₂ and MWCNTs Nanomicro Lubrication on the Process of 7050 Aluminum Alloy

Abstract

Nanofluid Minimum Quantity Lubrication (NMQL) is a resource-saving, environmentally friendly, and efficient green processing technology. Therefore, this study employs Minimum Quantity Lubrication (MQL) technology to conduct milling operations on aerospace 7050 aluminum alloy using soybean oil infused with varying concentrations of MoS₂ and MWCNTs nanoparticles. By measuring cutting forces, cutting temperatures, and surface roughness under three different lubrication conditions (dry machining, Minimum Quantity Lubrication, and nanofluid minimum quantity lubrication), the optimal lubricating oil with the best lubrication performance is selected. Under the conditions of hybrid nanofluid minimum quantity lubrication (NMQL), as compared to dry machining and Minimum Quantity Lubrication (MQL) processing, surface roughness was reduced by 48% and 36% respectively, cutting forces were decreased by 35% and 29% respectively, and cutting temperatures were lowered by 44% and 40%, respectively. Under the conditions of hybrid nanofluid minimum quantity lubrication, the optimal parameter combination is cutting speed (Vc) of 199.93 m/min, feed rate (f) of 0.18 mm, cutting depth (ap) of 0.49 mm, and nanofluid mass fraction (wt) of 0.51%. The hybrid nanofluid can significantly enhance heat exchange capacity and lubrication performance, thereby improving machining characteristics.

1. Introduction

Milling is a commonly used processing technology in modern manufacturing. Aluminum alloy is known for its low density, high strength, high temperature resistance, and sturdy load-bearing capacity, and has become a key material for manufacturing important structural components in the aerospace industry [1,2]. During the milling process, accompanied by intense friction and heat release, if not cooled in a timely manner, it can cause local temperatures to be too high, resulting in thermal damage to the surface of the workpiece [3]. Traditional pouring cooling processing of aluminum alloys often leads to problems such as excessive cutting force, severe tool wear, and poor surface quality of the workpiece [4,5]. The chemical additives contained in traditional coolants not only pollute the environment, but also pose a threat to workers’ health. Therefore, finding an efficient and environmentally friendly cutting fluid and processing method has become particularly important.

As a near-dry processing technology, micro lubrication (MQL) can effectively protect the environment and worker health. According to reports, compared to traditional pouring lubrication, MQL can reduce human toxicity by 87% and slow down the global warming trend by 21%. Compared with dry-cutting methods, micro lubrication can improve the surface quality of workpieces [6,7,8,9]. Khettabi et al. investigated the MQL and dry milling properties of 7075, 6061, and 2024 aluminum alloys. They found that MQL milling can reduce cutting force and improve surface finish [10]. Micro lubrication uses environmentally friendly vegetable oil and animal oil as lubricants, which will not cause pollution to the environment. Due to the high temperature and high speed involved in processing certain materials, traditional vegetable and animal oils often undergo oxidation, thereby reducing the lubrication effect during high-speed and high temperature processing. According to reports, adding nanoparticles to basic animal and vegetable oils can further enhance the thermal conductivity of the base fluid, improve its lubrication ability, and do no harm to the environment [11].

Singh et al. elaborated on the research progress of nano-cutting fluids. Research has found that compared to dry and traditional wet lubrication machining, the use of MQL technology with the addition of nanoparticles can significantly reduce cutting force and cutting temperature, while improving surface quality [12]. Albert et al. added graphite nanoparticles to the base solution and subsequently ground D-2 tool steel. Compared to traditional MQL grinding, this reduces grinding force and improves surface quality [13]. Rahmati et al. compared the surface quality changes during the milling process of AI6061-T6 alloy and evaluated the effects of using mineral oil and adding different mass fractions of MOS₂ nanoparticles on lubrication performance. The research results show that lubricants containing the minimum amount of MOS₂ nanoparticles perform better in improving surface quality compared to mineral oils [14]. Anuj et al. configured multi-walled carbon nanotubes and Al nanoparticles into cutting fluids with 0.25%, 0.75%, and 1.2% mass fractions, and subsequently conducted AISI 304 stainless steel turning experiments. As the mass fraction increases, the lubrication performance of the cutting fluid gradually improves significantly. Compared to a single Al nanofluid, the mixed nanofluid has a certain reduction in tool wear and peak processing temperature, which are 11% and 27.36%, respectively [15]. Sharma prepared a nanofluid cutting fluid using carbon nanotubes (CNTs) and conducted turning experiments on AISID-2 steel. The results showed that the temperature in the cutting area decreased significantly [16]. Aref added different qualities of MoS₂ and CuO nanoparticles to rapeseed oil and soybean oil, and ground AISID-2 tool steel. The research results show that adding 4% CuO nanoparticles and 2% MoS₂ nanoparticles to soybean oil can reduce the normal force of soybean oil by 19% and the tangential force by 35% [17]. Guan et al. proposed the use of a composite filled with oleic acid inside multi-walled carbon nanotubes to prepare nanofluids. Compared to ordinary multi-walled carbon nanotubes, the composite filled with oleic acid has better heat transfer and wettability within the multi-walled carbon nanotubes [18].

Multi-walled carbon nanotubes (MWCNTs) are nanotubes formed by winding multiple layers of graphite sheets around the centerline, with a hollow inner cavity structure inside [19]. The thermal conductivity of multi-walled carbon nanotubes can reach 6 kW/(m·k), which inherently has good lubrication properties and can play a role similar to “micro bearings”. Therefore, multi-walled carbon nanotubes are suitable as additives for cutting fluids [20]. MOS₂ nanoparticles have a layered structure, formed by covalent bonding between sulfur and molybdenum atoms, forming a hexagonal crystal structure [21,22]. The MOS₂ molecular layers are connected to each other through weak van der Waals forces, which endow MOS₂ with flexible structural by sliding the molecular layer, it can transform to the dual surface to form a transfer film, endowing MOS₂ with the characteristic of reducing friction.

This article uses soybean oil as the base fluid, combined with multi-walled carbon nanotubes with excellent thermal conductivity and molybdenum disulfide nanoparticles with excellent lubrication performance, to prepare three types of nanofluids: MOS₂ nanofluids with mass fractions of 0.5 wt.%, 1.0 wt.%, 1.5 wt.%, and 2.0 wt.%, MWCNTs nanofluids, and MOS₂/MWCNTs mixed nanofluids. Nanofluid-assisted milling experiments were conducted on 7050 aluminum alloy to evaluate the effects of nanofluid on cutting force F, cutting temperature °C, and surface roughness Ra. A nanofluid with a significant lubrication effect was selected and 16 orthogonal milling experiments were conducted to obtain an empirical formula for cutting force. The multi-objective genetic algorithm and TOPSIS comprehensive evaluation method were used to optimize various parameters, providing important reference values for practical nanofluid micro lubrication milling machining.

2. Materials and Methods

2.1. Experimental Setup

Milling experiments were conducted on the RAMBAUDI RAMMATIC 800 machining center using reverse milling. During the experiments, an HM300 infrared thermal imager was employed to measure temperature variations during the machining process. The cutting forces during milling were measured and collected using a Kistler 9257B dynamometer (Kistler, Winterthur, Switzerland) and DEWESoftX3 software. In terms of roughness measurement, the TR200 surface roughness meter is used.

In the aviation industry, due to its low density, high strength, high temperature resistance, and robust load-bearing capacity, 7050 aluminum alloy is extensively utilized in the fabrication of critical structural components [23]. However, further exploration is still needed for the milling research of such materials. Therefore, this study selected test samples with dimensions of 170 mm × 100 mm × 60 mm (length × width × height) for investigation. During the processing, a 20 mm diameter tool holder and the APMT1135PDER MH650 blade produced by Fengliang CNC Tool Company (Hanzhong, China)were used, with a square head and a sharp angle of 75 degrees. A milling cutter has two teeth. The tool material is hard alloy. The Minimum Quantity Lubrication (MQL) milling process and the measurement of milling forces are shown in Figure 1a,b.

Table 1. 7050 chemical composition (mass fraction %).

Element	Al	Cr	Zr	Zn	Si	Fe	Mn	Mg	Ti	Cu
Component/%	margin	≤0.04	0.08–0.15	5.7–6.7	≤0.12	0–0.15	≤0.1	1.9–2.6	≤0.12	1.9–2.6

provides detailed information on the chemical composition of the 7050 aluminum alloy.

2.2. The Preparation of Nanofluids

Mineral oil and synthetic oil are usually used as cutting fluids during the machining process. In the long run, these oils pose a threat to the environment and human health. We should use sustainable and environmentally friendly vegetable oil as cutting fluid. This cutting fluid has a positive impact on water, soil, air, and even operators [24]. In recent years, in order to enhance lubrication and cooling performance during processing, nanoparticles have been incorporated into vegetable oils [25].

This experiment employed a two-step method to prepare the nanofluid [26]. The two-step method refers to the process of mixing nanoparticles and a dispersant in a certain proportion, followed by preparing the nanofluid through ultrasonic vibration in a sonicator. This method’s preparation process is simple and easy to operate. This experiment utilized multi-walled carbon nanotubes with an average diameter of 50 nm produced by Xuzhou Jiechuang Innovative Materials Technology Co., Ltd. (Xuzhou, China), and molybdenum disulfide nanoparticles with an average particle size of 50 nm produced by Hefei Qianguo New Materials Technology Co., Ltd. (Hefei, China) (Data provided by the seller). Figure 2 shows SEM images of multi-walled carbon nanotubes and molybdenum disulfide nanoparticles (provided by the supply company). Add MoS₂ nanoparticles and multi-walled carbon nanotubes (MWCNTs) into soybean oil. Different mass fractions of nanoparticles were mixed with soybean oil and a dispersant (C₂H₆O₄S). The mixture was stirred with a magnetic stirrer for 10 min, followed by ultrasonic dispersion using an ultrasonic cleaner for 30 min. Nanofluids of MoS₂ and MWCNTs were prepared with mass fractions of 0.5 wt.%, 1.0 wt.%, 1.5 wt.%, and 2.0 wt.%. Additionally, a mixed nanofluid of MoS₂ and MWCNTs was also prepared.

2.3. Experimental Design

The study is divided into two stages. In the first stage, 14 sets of single-factor experiments were conducted, and the evaluation criteria included surface roughness, cutting force, and machining temperature (

Table 2. Machining parameters in the first phase.

Machining Parameters	Values
Cutting Speed ($v_c$)	100 m/min
Feed ($f$)	0.12 mm
Depth of cut ($a_p$)	0.8 mm
Minimal quantity lubrication (MQL) flow rate	50 mL/h
MQL inlet pressure	0.5/Mpa
Machining length	100 mm
MQL nozzle angle	30°
Distance between nozzle and tool	50 mm

). By comparing each set of experiments, the most significant nanofluid performance was selected. The second stage involved 16 sets of orthogonal experiments aimed at optimizing the best combination of process parameters under different conditions (refer to

Table 3. Machining parameters in the second phase.

Factors	Control Parameters	Parameter Range
	Parameter Combinations	1	2	3	4
A	$v_c\left(\mathrm{m}/\min \right)$	60	90	120	150
B	$f\left(\mathrm{mm}\right)$	0.08	0.1	0.12	0.14
C	$a_p\left(\mathrm{mm}\right)$	0.1	0.2	0.3	0.4
D	$wt\left(\%\right)$	0.5	1	1.5	2

and

Table 4. $L_{16}\left(4^4\right)$ orthogonal array experimentation and observations.

No.	${\mathit{v}}_{\mathit{c}}\left(\mathbf{m}/\mathbf{min}\right)$	$\mathit{f}\left(\mathbf{mm}\right)$	${\mathit{a}}_{\mathit{p}}\left(\mathbf{mm}\right)$	$\mathit{w}\mathit{t}\left(\mathit{\%}\right)$	$\mathit{F}\left(\mathit{N}\right)$
1	1	1	1	1	121.51
2	1	2	2	2	153.89
3	1	3	3	3	178.55
4	1	4	4	4	190.65
5	2	1	2	3	140.09
6	2	2	1	4	136.55
7	2	3	4	1	179.53
8	2	4	3	2	175.04
9	3	1	3	4	147.14
10	3	2	4	3	154.43
11	3	3	1	2	113.85
12	3	4	2	1	140.36
13	4	1	4	2	133.54
14	4	2	3	1	132.20
15	4	3	2	4	127.63
16	4	4	1	3	123.16

). After each experiment, replace the blade with a new one for the next experiment.

2.4. Optimization Routine

2.4.1. Multi-Objective Optimization Model

Genetic algorithms possess strong global optimization capabilities. The multi-objective optimization is conducted using the cutting force and material removal rate models as objective functions. The multi-objective optimization problem can be described as in Equations (1) and (2):

$$\min \left[f_1\left(x\right),f_2\left(x\right),\dots \dots ,f_m\left(x\right)\right]$$

(1)

$$s.t.\left\{\begin{array}{l}lb\le x_1\le ub\\ Aeq*x=beq\\ A*x\le b\end{array}\right.$$

(2)

where $f_i\left(x\right)$ is the objective function, $x$ is the variable to be optimized, $lb$ and $ub$ are the upper and lower limits of the variable, $Aeq*x=beq$ is a linear equality constraint, and $Aeq*x\le beq$ is a linear inequality constraint. In multi-objective optimization, the optimization of one objective function often comes at the expense of the degradation of another objective function. Therefore, the optimal solutions obtained through multi-objective genetic algorithms are referred to as non-dominated solutions or Pareto optimal solutions.

2.4.2. Parameter Settings

Set the genetic algorithm evolution generations to 1000, termination threshold to 1 × 10⁻¹⁰, initial population size to 200, Pareto front coefficient to 0.15, and leave all other settings at their default values. $x_1$ is the cutting speed, $x_2$ is the feed rate per tooth, $x_3$ is the back feed, and $x_4$ is the mass fraction of the mixed nanofluid. The constraint conditions are shown in Equation (3):

$$s.t.\left\{\begin{array}{l}50\le x_1\le 200\\ 0.1\le x_2\le 0.2\\ 0.1\le x_3\le 0.5\\ 0.5\le x_4\le 2\end{array}\right.$$

(3)

2.4.3. TOPSIS Comprehensive Evaluation

Assuming there are j evaluation objects and evaluation indicators, the original data matrix is shown in Equation (4):

$$X=\left[\begin{array}{l}{x}_{11}{x}_{12}\cdots x_{1j}\\ x_{21}{x}_{22}\cdots x_{2j}\\ ⋮\hspace{1em}⋮\hspace{1em}\ddots ⋮\\ x_{i1}{x}_{i2}\cdots x_{ij}\end{array}\right]$$

(4)

Step 1. Positive normalization of indicators.

Due to the many types of evaluation indicators, the direction of optimization is also different. It is necessary to normalize each indicator to unify it into a high optimization index The processing method is shown in Equation (5):

$${\stackrel{\sim }{x}}_{ij}=\left\{\begin{array}{l}{x}_{ij},\mathrm{High}\mathrm{Priority}\mathrm{Indicators}\\ 1/x_{ij},\mathrm{Low}\mathrm{Priority}\mathrm{Indicators}\\ M/\left(M+\left|x_{ij}-M\right|\right),\mathrm{Neutral}\mathrm{Indicators}\end{array}\right.$$

(5)

Step 2. Constructing a normalized initial matrix.

Normalize the forward data to eliminate errors caused by different dimensions of each evaluation indicator. Divide each column element by the norm of the current column vector (Equation (6)):

$$z_{ij}=\frac{x_{ij}}{\sqrt{{\displaystyle \sum _{i=1}^n{x}_{ij}^2}}}$$

(6)

The normalization matrix Z is obtained from the normalized raw data (Equation (7)):

$$Z=\left[\begin{array}{l}{z}_{11}{\mathrm{z}}_{12}\cdots {\mathrm{z}}_{1j}\\ z_{21}{\mathrm{z}}_{22}\cdots {\mathrm{z}}_{2j}\\ ⋮\hspace{1em}⋮\hspace{1em}\ddots ⋮\\ z_{i1}{\mathrm{z}}_{i2}\cdots {\mathrm{z}}_{ij}\end{array}\right]$$

(7)

Step 3. Determine positive and negative ideal solutions.

After normalization, the cosine method is used to find positive ideal solutions and negative ideal solutions. The positive ideal solution $Z^{+}$ is composed of the maximum values of each column of elements 1 × j. The negative ideal solution $Z^{-}$ of a dimensional matrix is determined by the minimum value of each column element 1 × j-dimensional matrix composition (Equations (8) and (9)):

$$Z^{+}=\left[\max \left\{z_{11},z_{21},\cdots ,\left.z_{i1}\right\}\right.,\max \left\{\left.z_{12},z_{22},\cdots ,z_{i2}\right\}\right.,\cdots ,\max \left\{\left.z_{1j},z_{2j},\cdots ,z_{ij}\right\}\right.\right]=\left(Z_1^{+},Z_2^{+},\cdots ,Z_j^{+}\right)$$

(8)

$$Z^{-}=\left[\min \left\{z_{11},z_{21},\cdots ,\left.z_{i1}\right\}\right.,\min \left\{\left.z_{12},z_{22},\cdots ,z_{i2}\right\}\right.,\cdots ,\min \left\{\left.z_{1j},z_{2j},\cdots ,z_{ij}\right\}\right.\right]=\left(Z_1^{-},Z_2^{-},\cdots ,Z_j^{-}\right)$$

(9)

Step 4. Calculate the distance between each evaluation object and positive and negative ideal solutions (Equation (10)).

$$D_i^{+}=\sqrt{{\displaystyle \sum _j{\left(z_{ij}-Z_j^{+}\right)}^2}},D_i^{-}=\sqrt{{{\displaystyle \sum _j\left(z_{ij}-Z_j^{-}\right)}}^2}$$

(10)

Step 5. Calculate the closeness of each evaluation object to the positive ideal solution (Equation (11)).

$$C_i=\frac{D_i^{-}}{D_i^{+}+D_i^{-}}$$

(11)

Among them, $0\le C_i\le 1$, the closer to 1, the higher the closeness between the evaluated object and the ideal solution, indicating that the evaluated object is better.

3. Results and Discussion

3.1. Surface Roughness

Analyzing the surface smoothness of workpieces is crucial as it directly affects dimensional accuracy and mechanical performance. Three measurements were taken at the same interval and the average value was taken to analyze the surface roughness (Ra) under different processing environments. Figure 3 shows the changes in roughness under different processing conditions. In Figure 3, the surface roughness of dry processing is the highest, reaching 1.652 µm. Next is the micro lubrication processing of soybean oil, with a surface roughness of 1.354 µm. Compared to dry machining, using soybean oil micro lubrication for machining can improve the surface quality of parts by 18%. When the mass fraction of MWCNTs nanofluids is 2%, the surface roughness is the lowest, at 0.974 µm. Compared to dry machining, nMQL (MWCNTs) improved surface quality by 41%. When the mass fraction of MoS₂ nanofluid is 2%, the surface roughness is the lowest, at 0.898 µm. Compared to dry machining, nMQL (MoS₂) improves surface quality by 45%. Only one type of nanoparticle nanofluid microlubrication did not achieve the optimal lubrication effect. When the mass fraction of MoS₂/MWCNTs mixed nanofluids reaches 2%, the Ra value of surface roughness is the smallest, which is 0.867 µm. Compared with dry machining, the use of nMQL (MoS₂/MWCNTs) can improve surface quality by 48%.

Analyzing the experimental results revealed the following patterns:

(1) As shown in Figure 3, under the condition of micro lubrication with a single nanofluid fluid. As the mass fraction of MoS₂ and MWCNTs nanofluids increases, the surface quality after processing gradually improves. This is because nanofluids with high mass fraction typically exhibit better resistance to high pressure and high temperature. Helps maintain stable lubrication performance in extreme operating environments. To maintain the high quality of the surface. The Ra value obtained from multi-walled carbon nanotubes (MWCNTs) nanofluids is greater than that obtained from molybdenum disulfide (MoS₂) nanofluids. This indicates that compared to MOS₂, the surface quality obtained using MWCNTs nanofluids for micro lubrication machining is poor.

The reason for the above phenomenon is that MoS₂ nanoparticles and MWCNTs nanoparticles have different physical and shape characteristics, leading to different lubrication mechanisms. Figure 4 shows the molecular structure of MoS₂ nanoparticles and the structure of multi-walled carbon nanotubes. The layered structure of MoS₂ nanoparticles is a hexagonal crystal structure formed by covalent bonding between molybdenum and sulfur atoms. The MOS₂ molecular layers are interconnected through weak van der Waals forces, endowing MOS₂ with a flexible structure. By sliding the molecular layer, it can form a complete lubricating oil film in the processing area, which endows MOS₂ with the characteristics of reducing friction. However, multi-walled carbon nanotubes (MWCNTs) are nanotubes formed by wrapping multiple layers of graphite around the centerline, with a hollow inner cavity structure inside. MWCNTs have a large aspect ratio and high strength due to this structure. MWCNTs are prone to agglomeration in nanofluids due to this structure. This further damages the integrity of the lubricating oil film on the surface of the friction pair, leading to an increase in the friction coefficient.

(2) According to the results, the MoS₂/MWCNTs mixed nanofluid obtained the lowest Ra value among the three nanofluids. So mixing nanofluids can achieve the best surface quality of the workpiece. This is because the mixed nanoparticles play a coating role in the lubrication mechanism (Figure 5). On the one hand, the coating phenomenon between MoS₂ nanoparticles and MWCNTs enhances the dispersion stability of MWCNTs in micro-lubricating cutting fluids. This encapsulation effect allows these two nanoparticles to leverage their respective lubricity advantages. The low shear plane of MOS₂ nanoparticles plays a role in reducing friction. The high thermal conductivity of MWCNTs nanoparticles prevents the rupture of the lubricating oil film under high temperature conditions. On the other hand, two different types of nanoparticles are fused into the mixed nanofluid, forming a polycrystalline protective film. Two types of nanoparticles complement each other and jointly protect the surface. This reduces friction and reduces surface roughness [27,28].

3.2. Cutting Temperature

Effectively controlling the generated heat during the machining process is crucial, as it directly impacts the tool lifespan, dimensional accuracy, and surface characteristics of the machined workpiece. The cutting temperature generated during the machining process was measured, as shown in Figure 6. As anticipated, under dry-cutting conditions without any lubricant, the recorded maximum temperature reached 187 °C. Under minimal lubrication milling conditions, the temperature peak of soybean oil reached a maximum of 172 °C. When the mass fraction of MWCNTs nanofluids is 2%, the temperature peak is the lowest at 113 °C. Compared with dry processing, nMQL (MWCNTs) reduces the temperature peak by 39.5%. When the mass fraction of MoS₂ nanofluid is 2%, the temperature peak is the lowest at 115 °C. Compared with dry processing, nMQL (MoS₂) reduces the temperature peak by 38.5%. The MoS₂/MWCNTs nanofluid mixture with a mass fraction of 2.0% reached a minimum of 104 °C, significantly reducing the peak temperature by 44% compared to dry-cutting conditions.

Analyzing the experimental results revealed the following patterns:

(1) As shown in Figure 6, under the condition of micro lubrication with a single nanofluid fluid. The peak processing temperature of MoS₂ nanofluids is higher than that of MWCNTs nanofluids. This is because compared to MWCNTs, MoS₂ has a lower thermal conductivity. Therefore, under minimum lubrication conditions, the temperature peak of MoS₂ nanofluid is higher than that of MWCNTs nanofluid.

(2) The MoS₂/MWCNTs nanofluid mixture with a mass fraction of 2.0% reached its lowest temperature peak (104 °C). There are two main reasons for improving the heat transfer performance of mixed nanofluids: First, the hybrid nanofluid improves the lubrication performance of the minimum lubrication processing zone through a synergistic coating mechanism. The acute Brownian motion between two types of nanoparticles enhances the heat transfer ability of the cutting area, thereby significantly reducing the temperature of the cutting area [25]. Second, the mixed nanofluid benefits from the simultaneous presence of MoS₂ and MWCNTs nanoparticles. Its heat transfer performance exceeds the minimum lubrication and the individual MoS₂ and MWCNTs nanofluids under minimum lubrication, thereby comprehensively improving the heat transfer ability of the cutting fluid. This means that mixed nanofluids under minimal lubrication are more effective in conducting and dispersing the heat generated during the cutting process by quickly dispersing heat into the fluid. As a result, the heat transferred to the workpiece decreases, resulting in a decrease in the peak milling temperature on the surface of the workpiece.

3.3. Cutting Force

Cutting force plays an important role in affecting cutting heat, tool wear, machining accuracy, and surface quality. Cutting force can reflect the lubrication performance of various cutting fluids. The higher the lubrication efficiency, the smaller the cutting force. Hence, studying the cutting forces for 7050 aluminum alloy is of paramount importance. The cutting forces for each concentration are illustrated in Figure 7. The maximum cutting force occurs during dry machining, measuring 374.42 N. When the mass fraction of MWCNTs nanofluids is 2%, the cutting force is the lowest, at 265.55 N. Compared with soybean oil minimum lubrication and dry processing, the cutting force was reduced by 23% and 29%, respectively. When the mass fraction of MoS₂ nanofluid is 2%, the cutting force is the lowest, 253.17 N. Compared with soybean oil minimum lubrication and dry processing, the cutting force was reduced by 26% and 32%, respectively. The MoS₂/MWCNTs nanofluid mixture with a mass fraction of 2.0% reached a minimum of 242.93 N. Compared with soybean oil minimum lubrication and dry processing, the cutting force is reduced by 29% and 35%, respectively. The partial diagram of cutting force is shown in Figure 8.

Analyzing the experimental results revealed the following patterns:

(1) As shown in Figure 7, MWCNTs nanofluid micro lubrication has a greater cutting force than MoS₂ nanofluid micro lubrication. This is because the molecular structure of MWCNTs lacks effective lubricating groups, and their chemical properties are stable. It is not easy to react in the cutting area, forming a relatively complete lubricating oil film. Therefore, under the same conditions, the cutting force is greater than that of MOS₂ nanofluids that can stably form and complete the oil film.

(2) From the results, it can be seen that the MoS₂/MWCNTs mixed nanofluid maximizes the reduction in cutting force. These two different types of nanoparticles can cross and combine to form a stronger and more stable lubricating oil film than a single nanofluid, enhancing the durability of the lubrication effect. These oil films can reduce the contact between the workpiece and the tool, thereby reducing cutting force. As the concentration of nanoparticles increases, they have more particles penetrating the gaps on the surface of the workpiece, shearing with the nanoparticles below to form a thicker lubricating film. This provides better protection for the tool and reduces cutting force. However, with the addition of a large number of nanoparticles, the aggregated nanoparticles may collide with the rough surface of the workpiece, affecting its surface quality and leading to higher milling forces [29]. This phenomenon explains why appropriate concentrations of nanoparticles are needed to effectively reduce milling forces.

The experimental results indicate that, compared to individual nanoparticles, the mixed nanofluid exhibits superior lubricating performance. To further explore the impact of mixed nanofluids on processing performance under different processing parameters, 16 sets of Taguchi orthogonal experiments were conducted (

Table 5. Significance test of cutting force empirical mode.

	Model	Sum of Square	Degree of Freedom	Sum of Mean Squares of Deviations	F-Value	${\mathit{R}}^2$
F	regression	0.333	4	0.083	33.146	0.923
	residual	0.028	11	0.003
	total	0.361	15

and

Table 6. The set of non-inferior solutions for region B.

Non-Inferior Solution	${\mathit{v}}_{\mathit{c}}(\mathbf{m}/\mathbf{min})$	$\mathit{f}(\mathbf{mm}/\mathit{z})$	${\mathit{a}}_{\mathit{p}}(\mathbf{mm})$	$\mathit{w}\mathit{t}(\%)$	$\mathit{F}(\mathit{N})$	$\mathit{Q}({\mathbf{cm}}^3/\mathbf{min})$
1	199.87	0.13	0.49	0.52	146.21	7.85
2	199.82	0.15	0.44	0.53	149.24	8.45
3	199.95	0.16	0.46	0.56	152.29	9.21
4	199.84	0.12	0.27	0.52	128.41	4.20
5	199.52	0.15	0.14	0.59	118.06	2.66
6	199.83	0.11	0.37	0.52	134.89	5.37
7	199.59	0.12	0.24	0.59	126.95	3.86
8	199.52	0.13	0.21	0.58	124.09	3.47
9	199.76	0.12	0.35	0.58	134.27	5.14
10	199.93	0.18	0.49	0.51	158.56	11.20
11	199.86	0.18	0.45	0.53	156.81	10.52
12	199.79	0.12	0.47	0.52	143.66	7.22
13	199.49	0.11	0.27	0.59	125.40	3.70
14	199.72	0.14	0.43	0.57	145.92	7.53
15	199.99	0.16	0.47	0.57	153.09	9.41
16	199.88	0.18	0.47	0.58	157.65	10.65

). The processing parameters are provided by the tool manufacturer. The micro lubrication process parameters are the same as the first stage experiment.

4. Parameter Optimization and Validation

4.1. Modeling

This study adopts a multiple linear regression model. Equation (12) provides an exponential regression model for cutting force.

$$F=c{v}_c^{b_1}{f}^{b_2}{a}_p^{b_3}{R}^{b_4}$$

(12)

where $F$ is the cutting force, $R$ is the mass fraction $wt$ of the mixed nanofluid, $v_c$ is the cutting speed, $f$ is the feed rate, $a_p$ is the cutting depth, $c$. $b_1$, $b_2$, $b_3$, $b_4$ are constants.

Take the logarithm on both sides of Formula (12) to obtain Formula (13). Convert Formula (13) to the linear expression in Formula (14). Use the multiple regression model formula for obtaining cutting force (15).

$$\ln F=\ln C+b_1\ln v_c+b_2\ln f+b_3\ln a_p+b_4\ln R$$

(13)

$$y=b_0+b_1{x}_1+b_2{x}_2+b_3{x}_3+b_4{x}_4$$

(14)

$$F=976.5247v_c^{-0.236}{f}^{0.23}{a}_p^{0.207}{R}^{0.037}$$

(15)

The significance test of the cutting force regression model was conducted using the F-value detection method. The significance test table for the empirical model of cutting force is shown in Table 5. In Table 5, it can be seen that the fitting degree of the empirical model of cutting force is very significant. It can better predict the milling force during the mixed nanofluid micro lubrication milling process. It can also provide theoretical guidance for the actual processing process.

4.2. Parameter Optimizations

Call the gamultiobj function for iterative optimization, and after reaching the termination condition threshold, obtain the Pareto optimal solution as shown in Figure 9.

From Figure 9, it can be seen that as the material removal rate increases, the cutting force gradually increases. The cutting force in area A is relatively high, which will affect the tool life. The material removal rate in region C is relatively low, which can affect processing efficiency. In region B, the cutting force and material removal rate remain within the expected range. Table 6 shows 1–16 non-inferior solutions selected from region B [30].

The TOPSIS evaluation method was used to evaluate the 16 non-inferior solutions in Table 6, and the evaluation results are shown in Figure 10. It can be found that the non-inferior solution 10 ranks first in the TOPSIS comprehensive evaluation method.

4.3. Verification Experiment

Finally, in order to test the reliability of the prediction results, experimental verification was conducted on the optimal solution (non-inferior solution 10), as shown in

Table 7. Comparison between predicted and experimental results.

Response Parameter	Predicted Value	Experimental Value	Deviation (%)
F (N)	158.56	169.34	6.8
Q (cm3/min)	11.20	11.68	4.3

. The maximum deviation (%) between the predicted experimental results and the actual experimental results does not exceed 6.8%, which verifies the rationality of the optimized results of the prediction model.

5. Conclusions

• The Minimum Quantity Lubrication (MQL) technology reduces cutting force, cutting temperature, and surface roughness due to the effectiveness of fuel injection pressure in the cutting area. In addition, in this method, adding nanoparticles to the base liquid can significantly improve lubrication efficiency compared to soybean oil.
• The lubrication performance of mixed nanofluids is the most significant. When the mass fraction is 2.0%, the cutting temperature and cutting force of MoS₂/MWCNTs mixed nanofluid during machining are 104 °C and 265.55 N, respectively. Compared with dry cutting, it has decreased by 44% and 35%, respectively. Surface roughness is 0.867 μm. Compared to dry cutting, it has increased by 48%.
• The TOPSIS method based on a genetic algorithm comprehensively evaluated the optimized combination parameters. The experiment proves that the deviation (%) between the predicted results and the experimental results does not exceed 6.8%. Therefore, it is recommended to combine these parameters to support the sustainable development of the manufacturing industry.