Analysis and Optimization of Friction Stir Welding Process for AA6092/ZrO₂ Composite Materials

Abstract

Aluminum metal matrix composites (AMMCs) are created by incorporating various ceramic particles into aluminum alloys, offering enhanced mechanical properties and customizable thermal and electrical characteristics. Zirconium dioxide (ZrO₂) is particularly valuable due to its excellent fracture toughness, wear resistance, low thermal conductivity, and high resistance to mechanical stress and cracking. Its compatibility with aluminum alloys makes it suitable for applications in medical products, electronic equipment, oxygen sensors, fuel cell membranes, and engine valve seats. Friction Stir Welding (FSW) is an emerging solid-state technique for joining AMMCs reinforced with particulate ceramics, effectively preserving their mechanical properties and reducing defects common in traditional fusion welding. This study aims to develop regression models that predict the Ultimate Tensile Strength (UTS), Percent Elongation (PE), and Weld Nugget Hardness (WNH) of friction stir welded AA6092 composites reinforced with ZrO₂ particles. The models will correlate with key FSW parameters, including tool rotational speed (TRS), Welding Speed (WS), Axial Load (AL), and the percentage of ZrO₂ reinforcement. Statistical software Design Expert, along with analysis of variance (ANOVA) and Student’s t-test, will be used to validate the developed models. Using Response Surface Methodology and ANOVA, optimized welding conditions were identified as TRS = 1208.09 rpm, WS = 51.5 mm/min, AL = 5.13 kN, and 15 wt.% ZrO₂. Under these parameters, the highest mechanical properties achieved were UTS = 494.81 MPa, PE = 8.51%, and WNH = 184.57 HRB. The developed regression models demonstrated high accuracy, with prediction errors below ±5% for all responses. These results confirm the effectiveness of the proposed models and parameters for enhancing the mechanical performance of AA6092/ZrO₂ FSW joints.

Keywords:
Friction stir welding; aluminium alloy AA6092; Zirconium dioxide; hardness; ultimate tensile strength; response surface methodology

1. Introduction

Reinforcing soft aluminum alloys with hard ceramic particles leads to aluminum metal matrix composites (AMMCs), which exhibit superior mechanical, thermal, and tribological properties. These enhancements have driven interest in applications across aerospace, automotive, marine, and military sectors. Various ceramic particles, such as SiO₂, TiO₂, and AlN_P, have been explored alongside traditional Al₂O₃ and SiC¹,². The ability to tailor properties by adjusting the reinforcement's weight percentage has increased their usage. AMMCs are particularly attractive due to their isotropic properties, higher operating temperatures, oxidation resistance, and easier fabrication compared to other reinforcement types³,⁴. While secondary processes like joining and machining are crucial, traditional fusion welding often results in reduced joint strength due to defects such as oxide inclusions and porosity. In contrast, solid-state welding techniques, like friction stir welding (FSW), offer a promising solution by avoiding these issues.

Friction Stir Welding (FSW) is an innovative joining technology for lightweight metals, developed by The Welding Institute (TWI) in 1991⁵,⁶. This method allows significant weight savings due to the high strength of FSW joints compared to conventional techniques. Operating below the melting point, FSW can join various materials, including aluminum, bronze, copper, and titanium, while producing less distortion and reducing costs. The process uses a non-consumable rotating tool with a specially designed pin and shoulder, which generates heat through friction and plastic deformation. As the tool moves, material is displaced from the front to the back of the pin, creating a strong joint through severe plastic deformation and dynamic recrystallization⁷,⁸. Material flow occurs in two modes: pin-driven, where material moves layer by layer, and shoulder-driven, where the shoulder deflects material for bulk movement. Insufficient axial pressure can lead to flash formation. FSW consists of three phases: plunge, main, and termination. In the plunge phase, the rotating tool is inserted into the joint, generating heat and displacing material around the pin. During the main phase, the tool travels along the joint, maintaining heat for sufficient softening. In the termination phase, the tool stops at the end of the joint and is withdrawn, leaving a keyhole that must be trimmed away⁹,¹⁰.

FSW is widely used for aluminum alloys, copper, magnesium, composite materials, polymers, and some ferrous materials, often resulting in a distinctive "onion-ring" structure in the stir zone. The thermo-mechanically affected zone (TMAZ) flanks the stir zone, exhibiting less strain and temperature impact, while the heat-affected zone (HAZ) undergoes thermal cycling without deformation, potentially affecting microstructure and mechanical properties, especially in age-hardened aluminum alloys¹¹-¹⁴. Moreover, in shipbuilding and marine industry, critical components such as panels for freeze, decks, oil rig panels, bulkheads and floors are made from various aluminum grades, with AA6092 currently serving as the industry standard for body panel/frame construction. While this alloy offers desirable properties like excellent machinability and formability, conventional fusion welding of aluminum alloys poses significant challenges. This welding method often leads to defects in the weld zone, including reduced strength, porosity, hot cracking, brittle solidification, an extensive heat-affected zone, and other discontinuities. In contrast, FSW present a viable alternative for materials that are difficult to fusion-weld, primarily due to their lower welding temperatures.

Research on the friction stir welding (FSW) of AA6092 composites is relatively scarce, with only a limited number of studies available in the literature. Dinaharan and Murugan¹⁵ investigated the effect of FSW process parameters like TRS, WS, AL and Wt.% of ZrB₂ on sliding wear Behavior of AA6061/0-10 wt.% ZrB₂ composite butt joints. They concluded that the FS Welded joints fabricated at TRS of 1150 rpm, WS of 50 mm/min and AL of 6 kN yielded highest wear resistance characteristic. Palanivel et al.¹⁶ adopted the central composite face-centered factorial design technique for optimizing the FSW process parameters like TRS, WS, tool shoulder profile, rotational speed, and welding speed to predict the tensile strength of dissimilar friction stir welded AA6351–AA5083 joints. They reported that the tool shoulder profile had the most significant effect on UTS of the dissimilar joints considered followed by TRS and WS. Finally, they concluded that the FS welded joints produced using straight square pin profiled tool with better tensile strength of 244.53 MPa at TRS of 950 rpm, WS of 63 mm/min and axial force of 14.7 kN. Ashok Kumar and Murugan¹⁷ developed regression models to predict the UTS and PE of FS welded AA6061 matrix composites reinforced with AlN_p particles. The models were created by correlating key parameters such as TRS, WS, AL, and Wt. % of AlN_p reinforcement in the AA6061 matrix.

Mohamed et al.¹⁸ discussed the optimization of FSW process parameters such as TRS and WS for AA 6061 T651 butt joint to attain out maximum UTS and Hardness using Response surface methodology (RSM). They reported that the optimum parameters for a higher UTS (207 MPa) and Hardness (70.2 HRB) obtained at the TRS of 950 rpm and WS of 4.55 mm/s. Vijayan and Rao¹⁹ optimized FSW process parameters such as TRS, WS, AL, and pin shapes for welding between AA2024 and AA6061 using a RSM based grey relational analysis approach combined with an entropy measurement technique. They reported that the optimal parameters for achieving a maximum UTS of 141 MPa were a TRS of 1700 rpm, a WS of 60 mm/min, an AL of 6 kN, and a square-shaped pin. Faradonbeh et al.²⁰ conducted FSW on Al-B4C composites and analysed the influences of various TRS, WS, and pin geometries on microstructure and mechanical characteristics of Al-B4C composites. They found that when the tool travel speed was increased or tool travel speed was decreased, the spreading of B4C particles became gradually constant in the aluminum matrix. They reported that Al-2%B4C composites welded by square pin showed higher UTS and joint efficiency.

Acharya et al.²¹ conducted FSW on AA6092/17.5 SiCp and studied the tool wear rate and its effect on the mechanical properties like UTS and WNH of the weld joint. They conducted the three sets of welding using H13 tool steel material having taper cylindrical pin profile by changing TRS, and keeping WS, toot tilt angle fixed at 2 mm/s and 2° respectively. They reported that minimum tool wear was observed at a TRS of 2000 rpm. Pandiyarajan et al.²² conducted a study on the metallurgical and mechanical characterization of FS welded stir-cast AA6061/6%ZrO₂/2%C hybrid metal matrix composites (MMCs). They concluded that the fine grain structure achieved under optimal conditions like TRS of 850 rpm, AL of 5 kN, and WS of 48 mm/min resulted in a maximum WNH of 53 HRB.

Salih et al.²³ investigated the FSW of AA6092/SiC/17.5p-T6 composites to assess the effects of varying TRS and WS on the metallurgical and mechanical properties. They reported that the highest UTS and joint efficiency were achieved at a TRS of 1500 rpm and a WS of 100 mm/min. Singh et al.²⁴ investigated the influence of aluminum oxide (Al₂O₃) nanoparticles on the mechanical and microstructural properties of friction stir welded (FSW) AA6061-T6 aluminum alloy. Researchers incorporated varying amounts of Al₂O₃ nanoparticles into the weld zone via a groove-filling technique to create composite joints. It is evaluated that this reinforcement impacted tensile strength, elongation, and microhardness. Microstructural analysis revealed improved grain refinement and particle dispersion within the stir zone. The findings indicated that Al₂O₃ nanoparticles significantly enhance weld strength and hardness, though with a slight reduction in ductility.

Singh et al.²⁵ systematically optimized friction stir welding (FSW) parameters for producing AA6061-T6 aluminum matrix nanocomposites reinforced with Al₂O₃ nanoparticles. Utilizing the Taguchi design of experiments, the study evaluated how tool rotation speed, traverse speed, and axial force affect nanoparticle distribution and weld quality. Microstructural characteristics were investigated using various microscopy techniques, while hardness and tensile strength were assessed as mechanical properties. The optimized parameters achieved through this study resulted in superior dispersion of nanoparticles and significant improvements in the composite's mechanical performance. Singh et al.²⁶ investigated the mechanical and microstructural behavior of AA6061-T6 aluminum alloy joints fabricated by friction stir welding (FSW) with nano-sized reinforcement particles, aiming to enhance joint performance by forming aluminum matrix nanocomposites. Detailed mechanical testing, including tensile strength, microhardness, and elongation, was performed. Concurrently, optical and scanning electron microscopy evaluated grain refinement, particle distribution, and interfacial bonding. The results demonstrated that nanoparticle inclusion significantly improved weld hardness and strength, though some trade-offs in ductility were observed. Maneiah et al.²⁷ discussed the FSW of AA6061-T6 alloy using H13 tool steel hexagonal pin taking the FSW process parameters such as TRS, tilt angle, WS, on the UTS of the FSW joints. They validated the experimental condition using multiple regression optimizations of desirability functions. They reported that the highest UTS of 191 MPa at a TRS of 1400 rpm, tilt angle of 0º and WS of 100 mm/min. Singh²⁸ offered a comprehensive overview of recent advancements in developing nanoparticle-reinforced joints fabricated via friction stir welding (FSW). It examined incorporating various nanoparticles like Al₂O₃, SiC, TiO₂, and CNTs into aluminum and other metal matrix composites to enhance joint performance. The paper discussed how reinforcement type, size, and distribution influence crucial mechanical properties (e.g., tensile strength, hardness, wear resistance) and microstructural features (e.g., grain refinement, particle dispersion). Furthermore, it addressed challenges in achieving uniform nanoparticle distribution and strong interfacial bonding during FSW, highlighting relevant experimental techniques, parameter optimization, and post-weld treatments. Singh and Sharma²⁹ investigated the microstructural evolution and mechanical strengthening mechanisms in friction stir welded (FSW) AA6061-T6 joints reinforced with Al₂O₃ nanoparticles. The study aimed to understand how incorporating these ceramic nanoparticles influences grain refinement, dislocation behavior, and overall joint performance. Through detailed optical and scanning electron microscopy analyses, the authors demonstrated a significant grain size reduction in the stir zone, attributing it to dynamic recrystallization and effective nanoparticle dispersion. Mechanical testing revealed substantial improvements in tensile strength and hardness, primarily linked to mechanisms like Orowan strengthening, load transfer, and grain boundary pinning.

In recent years, substantial progress has been made in the study of Friction Stir Welding (FSW) of Aluminum Metal Matrix Composites (AMMCs). The maximum ZrO₂ content was limited to 20 wt.% to avoid particle agglomeration, poor wettability, and porosity issues. This upper limit is consistent with prior studies⁵,²¹,²², which demonstrate that aluminum matrix composites can accommodate up to 20 wt.% ceramic reinforcements with proper dispersion techniques. While extensive research has been carried out on the friction stir welding (FSW) of aluminum metal matrix composites (AMMCs), the majority of studies have focused on alloys such as AA6061 and AA2024 reinforced with ceramic particulates like Al₂O₃, SiC, and B₄C. However, there is a noticeable paucity of data concerning the FSW behavior of AA6092 reinforced with ZrO₂ particles, especially in terms of correlating process parameters with key mechanical properties such as Ultimate Tensile Strength (UTS), Percentage Elongation (PE), and Weld Nugget Hardness (WNH).

This research addresses that gap by presenting a comprehensive experimental study on FSW of stir-cast AA6092/ZrO₂ composites, which, to the best of the authors’ knowledge, has not been systematically investigated in prior literature. The novelty of the present work lies in the development and validation of statistical regression models using central composite rotatable design to optimize multiple FSW parameters simultaneously. Moreover, the study incorporates response surface methodology (RSM) and ANOVA to evaluate the individual and interaction effects of parameters—offering a predictive framework for joint performance metrics. This makes the current work both novel and practically valuable for industries aiming to adopt advanced joining methods for complex composite materials.

However, the literature still lacks comprehensive research on the FSW of AA6092/ZrO₂ composites, particularly regarding key aspects of joint performance such as Weld Nugget Hardness (WNH), Ultimate Tensile Strength (UTS), and Percent Elongation (PE). This study attempts to join Recording the response parametersthe AA6092/ZrO₂ composite using the FSW process. To optimize the number of experimental trials, a four-factor, five-level central composite rotatable design matrix was employed. Three regression models were developed to establish relationships between significant parameters—tool rotational speed (TRS), welding speed (WS), axial load (AL), and percentage of reinforcement—and the UTS, PE, and WNH of the friction stir welded AA6092/ZrO₂ composite. The regression models were then optimized using the generalized reduced gradient method to maximize UTS under three different conditions: (i) maximizing UTS, (ii) maximizing UTS at higher welding speeds, and (iii) maximizing UTS, PE, and WNH at higher welding speeds.

2. Scheme of Evaluation

2.1. Production of AA6092/ZrO₂ composite

The stir casting technique was employed for fabrication of AA6092/ZrO₂ composite. An indigenously developed modified electric stir casting furnace with a bottom-pouring arrangement was employed for composite fabrication. Cleaned extruded AA6092 rods, 25 mm in diameter, were placed into a coated stainless-steel crucible, and the furnace temperature was set to 1000°C. The chemical composition of the AA6092 alloy is detailed in Table 1. The melt was stirred using a coated stainless-steel stirrer driven by an electric motor to ensure the incorporation and uniform distribution of ZrO₂ reinforcement within the molten AA6092 alloy. To prevent contamination at high temperatures, both the crucible and stirrer were coated. A predetermined quantity of preheated ZrO₂ particles (3–4 µm in size) was added at the vortex's edge. To improve the wettability of the ZrO₂ particles with the AA6092 alloy matrix, magnesium amounting to 2 wt.% of the total composite weight was added. The ZrO₂ particles were incorporated into the melt for 260 seconds, after which the mixture was further stirred for 1200 seconds before being poured into a preheated permanent mold (100 mm × 50 mm × 50 mm) through the bottom-pouring system. Argon gas was supplied at a constant flow rate of 2 lpm once the furnace reached 650°C and continued until the molten composite was poured into the mold. Various AA6092/ ZrO₂ composites containing 0–20 wt.% of ZrO₂ were produced in this manner³⁰,³¹. Plates measuring 100 mm × 50 mm × 5 mm (shown in Figure 1) were cut from the composite block for FSW trials, process parameter optimization, and FSW experiments as per the design matrix. Figure 2(a) shows the SEM analysis of the stir-cast AA6092 matrix alloy, while Figures 2(b-e) present SEM images of AA6092/ ZrO₂ composites for samples 2 through 5. The SEM analysis clearly indicates that the ZrO₂ particles are uniformly distributed, embedded within the primary aluminum dendrites rather than concentrated in the inter-dendritic zones. While uniform particle dispersion is visible, clear grain refinement is not evident at this magnification; higher-resolution techniques are needed for grain size quantification."

Table 1
The chemical composition of AA6092 alloy matrix.

Element	Mg	Si	Fe	Mn	Cu	Cr	Zn	Ti	Al
Wt. %	1.2	0.4	0.3	0.15	0.7	0.15	0.25	0.15	Balance

Figure 1
The fabricated stir cast AA6092/ZrO₂ composite plate.

Figure 2
SEM analysis of AA6092 alloy and AA6092/ ZrO₂ composites:(a) Sample 1 (Casted AA6092), (b) Sample 2 (95% AA6092/5% AA6092/ZrO₂), (c) Sample3 (90% AA6092/10% AA6092/ZrO₂, (d) Sample 4 (85% AA6092/15% AA6092/ZrO₂), (e) Sample 5 (80% AA6092/20% AA6092/ZrO₂). ZrO₂.

The ultimate tensile strength (UTS) of the fabricated AA6092/ZrO₂ stir-cast composites was evaluated using three standard test specimens, prepared in accordance with ASTM E8-04, and tested on a computerized universal tensile testing machine. The Hardness of the composites was measured using three standard test specimens, prepared according to ASTM E10-08, and tested on a Rockwell Hardness Tester. Dry sliding wear tests were conducted using a pin-on-disc machine following the ASTM G99 standard. The average UTS, Hardness and wear rate values are provided in Table 2.

Table 2
UTS, hardness and wear rate of AA6092 composites.

Sample No		1	2	3	4	5
Wt.% of compositions	AA6092	100	95	90	85	80
	ZrO2	0	5	10	15	20
UTS (MPa)		448.7	468.48	483.7	492.56	480.56
% of variation		-	4.4	7.8	9.77	7.10
Hardness (HRB)		163	165.3	170.5	172.5	175.7
% of variation		-	1.41	4.60	5.83	7.79
Wear Rate (1 x e -7 g/mm)		2.58	1.82	1.63	1.32	1.35
% of variation		-	29.46	36.82	48.84	47.67

2.2. Process parameters identification

Based on preliminary trials and literature reviews, several independent FSW process parameters affecting ultimate tensile strength (UTS), percent elongation (PE) and weld nugget Hardness (WNH) were identified, including tool pin profile (P), tool rotational speed (TRS), welding speed (WS), and axial load (AL). The primary FSW parameters influencing joint properties are tool rotational speed (TRS), welding speed (WS), and axial load (AL). Additionally, the tool pin profile significantly impacts joint characteristics, A square pin-profile tool made from High Carbon High Chromium (HCHCr) steel, oil-hardened to 62 HRC, was used. The tool dimensions included a shoulder diameter of 18 mm, pin length of 5.7 mm, and pin diameter of 4.24 mm. The geometry was finalized based on literature benchmarks and initial trials to ensure defect-free joints. The square profile enhances mechanical intermixing and material flow compared to cylindrical and threaded profiles (Figure 3). The chemical composition of this HCHCr steel is represented in Table 3. These square pin profiles were fabricated using Computerized Numerically Controlled (CNC) turning centres and Electrical Discharge Machines (EDM). Consequently, the weight percentage of ZrO₂ particles (W) was also considered as a factor to understand the effect of weight percentages of ZrO₂ particles on UTS, PE, and WNH.

Figure 3
(a) 2-D drawing of FSW tool, (b) CAD model of FSW tool (details of FSW tool geometry).

Table 3
Chemical composition of high carbon high chromium steel.

Element	C	Mn	Si	Co	Cr	Mo	V	P	Ni	Cu	S
Wt. %	1.4-1.6	0.6	0.6	1	11-13	0.7-1.2	1.1	0.03	0.3	0.25	0.03

2.3. Limits of FSW process parameters identification

The range of process parameters within which friction stir welding can be performed without defects—such as voids, pinholes, wormholes, tunnels of varying sizes, and cracks—is commonly referred to as the welding window. The friction stir welding (FSW) window for achieving sound welds in aluminum metal matrix composites (AMMCs) is narrower than that for unreinforced alloys due to the presence of ceramic particles. Additionally, the reduced ductility of cast composites further constricts the FSW window. To establish the working ranges for all selected factors, a series of trial welds were conducted, and cross-sectional examinations were performed to identify any defects in the weld zone. The limits for each factor were determined to ensure defect-free welds. For data recording and processing convenience, the upper limit of a factor was assigned a code of +2, while the lower limit was coded as -2. Intermediate values were calculated using the following Equation 1³²,³³:

$$Z_i=2\left[2Z-\left(Z_{max}+Z_{min}\right)\right]/(Z_{max}-Z_{min})$$ (1)

where Z_i is the essential coded value of a variable Z; Z is any value of the variable between Z_max and Z_min; Z_max is the upper limit of the variable; Z_min is the lower limit of the variable. The selected levels of the process parameters with their notations and units are depicted in Table 4.

Table 4
Process parameters and its levels.

S. No.	Process Parameter	Notation	Unit	Levels
				-2	-1	0	1	2
1.	Tool Rotational Speed	TRS	rpm	1000	1100	1200	1300	1400
2.	Welding Speed	WS	mm/min	25	40	55	70	85
3.	Axial Load	AL	kN	3	4	5	6	7
4.	% of ZrO2	W	Wt.%	0	5	10	15	20

2.4. Developing an experimental design matrix

The experimental design matrix selected, as shown in Table 4, is a central composite rotatable full factorial design comprising of 31 sets of coded conditions. The initial 16 experimental runs are based on a full factorial design matrix (2⁴=16). Following this, the next 8 runs involve combinations of each process variable. The centre points are defined by all variables at the intermediate (0) level, while the star points consist of combinations of each variable at either its lowest (-2) or highest (+2) level, with the other three variables remaining at intermediate levels. The remaining 7 experimental runs have all variables set at the intermediate level (0), representing the centre points. Consequently, these 31 experimental runs enable the estimation of linear, quadratic, and two-way interactive effects of the process parameters on the responses.

2.5. Conducting the experiments as per the design matrix

The experiments were conducted following the design matrix (Table 4) using a square butt joint configuration sized 100 mm × 50 mm × 6 mm (AA6092/ZrO2 composite plates), employing a single-pass butt welding procedure. Prior to friction stir welding, surface oxides on the plates were removed through wire brushing, and the welding direction was aligned with the plate's rolling direction. To minimize the impact of unknown nuisance variables, thirty-one weld runs were executed randomly from the design matrix. A semi-automatic friction stir welding (FSW) machine (M/s RV Machine Tools, Coimbatore, INDIA) was used, with tool rotational speed and welding speed adjusted as per each test run. The tool pin was inserted into the abutting surfaces until the shoulder made contact, and the specified axial load was applied to exert pressure during the runs. After a brief dwell period to generate sufficient frictional heat for plasticizing the material, the FSW machine table moved at a constant speed as outlined in the design matrix. The axial load was maintained throughout each run, ensuring consistent conditions for all subsequent welding operations. A typical experimental setup of FSW machine is illustrated in Figure 4.

Figure 4
FSW Experimental set up.

2.6. Recording the response parameters

Three tensile specimens were prepared from each of the thirty-one welded plates by cutting them perpendicular to the welding direction to the specified dimensions, in accordance with ASTM E8M-04 standards. The Ultimate Tensile Strength (UTS) and Percentage Elongation (PE) of the specimens were measured at room temperature using a Computerized Universal Testing Machine (Make: Associated Scientific Engineering Works, Model: F-100, Capacity: Max 5 Ton). The average UTS and PE values for the base composite and friction stir welded composite joints are presented in Tables 2 and 5, respectively. Table 6 represents standard deviation to UTS, PE, JE and WNH. The joint efficiency of the welded composites was calculated using Equation (2), with the joint efficiency of all thirty-one welded joints detailed in Table 5.

Table 5
Design matrix and experimental results.

Run	Tool rotational speed (rpm)	Welding speed (mm/min)	Axial force (kN)	% of ZrO2 (Wt.%)	UTS (MPa)	PE (%)	Joint efficiency (%)	WNH (HRB)
1	1200	55	5	20	540.14	9.56	93.9413	195.5
2	1200	55	5	10	494.24	8.27	90.4137	183
3	1200	55	3	10	446.52	6.52	81.4606	165.5
4	1200	55	5	0	467.94	6.96	93.6975	168
5	1300	40	4	15	488.87	7.94	87.1307	179
6	1200	55	7	10	474.05	7.52	86.6257	175
7	1300	40	6	5	458.68	6.42	88.8825	161.4
8	1300	70	4	15	472.59	7.26	84.1545	175
9	1100	40	6	5	459.65	6.68	89.0757	167
10	1100	70	4	15	472.32	7.09	84.1051	175
11	1200	55	5	10	493.26	8.22	90.2298	183
12	1200	55	5	10	495.45	8.42	90.6407	185
13	1200	85	5	10	445.08	4.98	81.1904	151
14	1200	55	5	10	495.86	8.97	90.7176	185
15	1100	70	6	5	453.64	5.85	87.8785	158.5
16	1100	70	4	5	448.2	5.64	86.7948	158
17	1000	55	5	10	457.33	6.19	83.4887	160.2
18	1100	40	4	5	465.01	6.93	90.1434	168
19	1100	40	6	15	489.18	8.01	87.1874	179
20	1200	55	5	10	493.26	8.22	90.2298	183
21	1300	70	4	5	461.21	6.9	89.3865	167
22	1300	70	6	5	456.86	5.91	88.5199	159
23	1200	25	5	10	458.88	6.6	83.7795	166
24	1400	55	5	10	471.27	7.08	86.1041	174
25	1200	55	5	10	495.45	8.42	90.6407	184
26	1100	70	6	15	467.64	6.96	83.2495	170.1
27	1300	40	6	15	482.35	7.61	85.9388	177
28	1300	40	4	5	474.13	7.54	91.9602	176
29	1300	70	6	15	486.06	7.62	86.617	178
30	1200	55	5	10	496.12	9.07	90.7664	189
31	1100	40	4	15	489.53	8.15	87.2514	179

Table 6
Standard deviation to UTS, PE, JE and WNH.

Property	Standard Deviation
Ultimate Tensile Strength (UTS)	±20.38 MPa
Percentage Elongation (PE)	±1.07%
Weld Nugget Hardness (WNH)	±10.47 HRB
Joint Efficiency (JE)	±3.54%

$$JointEfficiency\left(\%\right)=\left(UT{S}_{welded}{}_{joint}/UT{S}_{composite}\right)x100$$ (2)

Weld nugget Hardness testing was performed using a Rockwell Hardness Tester (Wilson Wolpert, Germany). Test specimens were prepared in accordance with ASTM E10-08 standards. Hardness measurements were taken at three locations: one in the weld nugget zone (WNZ) and two in the heat-affected zone (HAZ). Metallographic samples were obtained from the transverse section of the welded plate and treated with a colour etchant solution composed of 4 g of KMnO₄ and 1 g of NaOH dissolved in 100 ml of distilled water. The metallurgical structures of the samples were analysed using an optical metallurgical microscope (De-Wintour Inverted Trinocular Metallurgical Microscope) and a scanning electron microscope (FEI SEM-Apreo Model).

3. Developing Empirical Relationships

The response variables like Ultimate Tensile Strength (UTS), Percentage Elongation (PE), and Weld Nugget Hardness (WNH) in the Friction Stir Welding (FSW) of AA6092 composites are functions of Tool Rotational Speed (TRS), Welding Speed (WS), Axial Load (AL), and the percentage of Zirconium dioxide (ZrO₂) reinforcement (W). The corresponding surfaces for these relationships are represented in Equations (3), (4), and (5), respectively³⁴,³⁵.

$$UTS=f\left(TRS,WS,AL,W\right)$$ (3)

$$PE=f\left(TRS,WS,AL,W\right)$$ (4)

$$WNH=f\left(TRS,WS,AL,W\right)$$ (5)

The second-order polynomial (regression) Equation (6) employed to denote the response surface is expressed by

$$X=a_0+a_i{x}_i+a_{ii}{x}_i{}^2+a_{ij}{x}_i{x}_j+\epsilon$$ (6)

A second-order polynomial regression model is used to represent the response surface for the variable ‘X’. In this model, a₀ is the average response; a_i, a_ii and a_ij are coefficients reliant on the major and interaction influences of the parameters; and ε is the statistical error. The coefficients were calculated using Design Expert software and tested for significance at a 95% confidence level. Insignificant terms were removed from the model without sacrificing its accuracy, thereby simplifying the regression process. The finalized regression models for predicting Ultimate Tensile Strength (UTS), Percentage Elongation (PE), and Weld Nugget Hardness (WNH) in friction stir welded joints are presented in coded form, incorporating only the significant control parameters. The final regression equations for computing UTS, PE, and WNH in the FSW of AA6092 composites are shown in Equations (7), (8), and (9), respectively. Tables 7, 8, and 9 provide the ANOVA results for UTS, PE, and WNH.

Table 7
ANOVA results for Ultimate Tensile Strength (UTS).

Source	Sum of Squares	df	Mean Square	F-value	p-value
Model	11633.14	14	830.94	16.01	< 0.0001	significant
A-Tool Rotational Speed	167.80	1	167.80	3.23	0.0911
B-Welding Speed	565.32	1	565.32	10.89	0.0045
C-Axial Load	57.85	1	57.85	1.11	0.3068
D-% of ZrO2	4149.09	1	4149.09	79.92	< 0.0001
AB	73.36	1	73.36	1.41	0.2519
AC	3.90	1	3.90	0.0751	0.7875
AD	10.86	1	10.86	0.2091	0.6536
BC	88.17	1	88.17	1.70	0.2109
BD	11.83	1	11.83	0.2279	0.6395
CD	29.27	1	29.27	0.5638	0.4637
A2	1647.33	1	1647.33	31.73	< 0.0001
B2	3255.58	1	3255.58	62.71	< 0.0001
C2	2111.85	1	2111.85	40.68	< 0.0001
D2	157.25	1	157.25	3.03	0.1010
Residual	830.66	16	51.92
Lack of Fit	821.89	10	82.19	1.25	< 0.0001	significant
Pure Error	8.77	6	1.46
Cor Total	12463.80	30

Table 8
ANOVA results for Percentage Elongation (PE).

Source	Sum of Squares	df	Mean Square	F-value	p-value
Model	32.12	14	2.29	16.45	< 0.0001	significant
A-Tool Rotational Speed	0.5612	1	0.5612	4.02	0.0621
B-Welding Speed	3.60	1	3.60	25.78	0.0001
C-Axial Load	0.0063	1	0.0063	0.0454	0.8339
D-% of ZrO2	8.13	1	8.13	58.30	< 0.0001
AB	0.3630	1	0.3630	2.60	0.1262
AC	0.1958	1	0.1958	1.40	0.2534
AD	0.1314	1	0.1314	0.9421	0.3462
BC	0.1040	1	0.1040	0.7457	0.4006
BD	0.0150	1	0.0150	0.1076	0.7472
CD	0.2280	1	0.2280	1.63	0.2193
A2	5.84	1	5.84	41.86	< 0.0001
B2	12.57	1	12.57	90.15	< 0.0001
C2	3.62	1	3.62	25.93	0.0001
D2	0.0595	1	0.0595	0.4265	0.5230
Residual	2.23	16	0.1395
Lack of Fit	1.46	10	0.1465	1.15	0.4534	not significant
Pure Error	0.7671	6	0.1279
Cor Total	34.35	30

Table 9
ANOVA results for Weld Nugget Hardness (WNH).

Source	Sum of Squares	df	Mean Square	F-value	p-value
Model	3060.29	14	218.59	15.39	< 0.0001	significant
A-Tool Rotational Speed	85.88	1	85.88	6.05	0.0257
B-Welding Speed	239.40	1	239.40	16.86	0.0008
C-Axial Load	2.67	1	2.67	0.1878	0.6706
D-% of ZrO2	965.20	1	965.20	67.96	< 0.0001
AB	18.06	1	18.06	1.27	0.2760
AC	16.40	1	16.40	1.15	0.2984
AD	2.25	1	2.25	0.1584	0.6959
BC	4.20	1	4.20	0.2959	0.5940
BD	12.25	1	12.25	0.8626	0.3668
CD	23.04	1	23.04	1.62	0.2210
A2	503.21	1	503.21	35.43	< 0.0001
B2	1151.22	1	1151.22	81.06	< 0.0001
C2	332.02	1	332.02	23.38	0.0002
D2	8.11	1	8.11	0.5708	0.4609
Residual	227.22	16	14.20
Lack of Fit	199.51	10	19.95	4.32	0.0434	significant
Pure Error	27.71	6	4.62
Cor Total	3287.51	30

UTS, MPa = - 899.9825 + 1.81068 x TRS + 2.51204 x WS + 82.1022 x AL + 2.00881 x W + (0.001427 x TRS x WS) - (0.004938 x TRS x AL) - (0.001648 x TRS x W) + (0.1565 x WS x AL) - (0.011467 x WS x W) + (0.2705 x AL x W) - (0.000759 x TRS ²) - (0.047422 x WS ²) - (8.59372 x AL ²) + (0.093801 x W ²) (7)

PE, % = -75.320 + 0.1118 x TRS + 0.1469 x WS + 4.3330 x AL + 0.2286 x W + (0.0001 x TRS x WS) - (0.00111 x TRS x AL) - (0.0002 x TRS x W) + (0.0054 x WS x AL) + (0.0004 x WS x W) + (0.02388 x AL x W) - (0.000043 x TRS ²) - (0.0029 x WS ²) - (0.3556 x AL ²) - (0.0018 x W ²) (8)

WNH, HRB = -609.4436 + 1.0449 x TRS + 1.7539 x WS + 41.6119 x AL + 0.7526 x W + (0.0007 x TRS x WS) - (0.0101 x TRS x AL) - (0.0008 x TRS x W) + (0.03417 x WS x AL) + (0.0117 x WS x W) + (0.240 x AL x W) - (0.0004 x TRS ²) - (0.0282 x WS ²) - (3.4074 x AL ²) - (0.0214 x W ²) (9)

Table 7 presents the ANOVA results for Ultimate Tensile Strength (UTS). A 95% confidence level and a 5% significance level were used to assess the significance of each term in the regression model. The analysis identified three key factors significantly influencing Ultimate Tensile Strength (UTS): the Percentage of ZrO₂ (most significant), the Welding Speed (second-highest influence), and Tool Rotational Speed (third-highest influence). These factors were deemed significant based on their p-values, all of which were below 0.05. Additionally, a lack-of-fit test was conducted to evaluate the adequacy of the regression model. The calculated F-value for lack of fit was 1.25, substantially lower than the standard F-value of 16.01 at the 95% confidence level, indicating that the model provides a reliable fit for predicting Ultimate Tensile Strength (UTS) within the specified range of FSW process parameters and their levels.

Table 8 presents the ANOVA results for Percentage Elongation (PE). Using the same confidence and significance levels, the analysis revealed that Percentage Elongation (PE) is significantly influenced by the Percentage of ZrO₂ (greatest contribution), the Welding Speed (second-highest influence), and Tool Rotational Speed (third-highest influence). These factors were found significant based on their p-values, all below 0.05. The lack-of-fit test further validated the model's adequacy, with a calculated F-value of 1.15—substantially lower than the standard F-value of 16.45 at the 95% confidence level. This confirms that the regression model is a reliable predictor of Percentage Elongation (PE) within the given range of FSW process parameters and levels.

Table 9's ANOVA for Weld Nugget Hardness (WNH) at a 95% confidence level identified key influencing factors. The percentage of ZrO₂ showed the greatest impact, followed by the Welding Speed, and then Tool Rotational Speed itself, all with p-values below 0.05. A lack-of-fit F-value of 4.32, significantly less than the critical 15.39, confirms the regression model's adequacy for predicting Weld Nugget Hardness (WNH) within the studied FSW process parameter ranges.

4. The Effect of the FSW Process Parameters on the Responses (UTS, PE and WNH)

4.1. Ultimate tensile strength (UTS)

This section analyses the influence of FSW process parameters on three response variables: Ultimate Tensile Strength (UTS), Percentage Elongation (PE), and Weld Nugget Hardness (WNH). The interaction effects of two input process parameters on these responses were examined, while keeping the third process parameter constant at its average level. For UTS, the perturbation plot (Figure 5) illustrates the impact of the FSW process parameters within an optimized design framework. The plot shows how UTS changes as each parameter deviates from its reference point, with all other parameters held constant. The results indicate that Tool Rotational Speed (TRS) is the most influential parameter affecting UTS, followed by Welding Speed (WS) and Axial Load (AL), in order of decreasing significance. Additionally, the plot reveals that the weight percentage of ZrO₂ (Wt. % ZrO₂) has the greatest impact on UTS, followed by TRS, AL, and WS.

Figure 5
Perturbation plot (impact of process parameters on the UTS).

The interaction effects of Tool Rotational Speed (TRS), Welding Speed (WS), Axial Load (AL), and ZrO₂ content (W) on Ultimate Tensile Strength (UTS) are shown in Figures 6, 7 and 8. Figures 6(a) and 6(b) depict the influence of TRS and WS on UTS, while keeping AL and W constant at 5 kN and 10 wt.%, respectively. The 2D contour plot in Figure 6(a) shows concentric circles representing tensile strength (in MPa), with the optimal UTS value located at the centre of the plot. Specifically, the optimum UTS of 494.8 MPa is achieved at a TRS of 1200 rpm and a WS of 55 mm/min, as shown in both Figures 6(a) and 6(b). Figure 7 demonstrates the interaction effect of AL and TRS on UTS, while WS and W are held constant at 55 mm/min and 10 wt.%, respectively. The optimal UTS of approximately 494.8 MPa is achieved at a TRS of 1200 rpm and an AL of 5 kN, as shown in both the 2D and 3D plots.

Figure 6
(a) 2D contour plot, (b) 3D contour plot (effect of TRS and WS on UTS of FSW joint.

Figure 7
(a) 2D contour plot, (b) 3D contour plot (effect of TRS and AL on UTS of FSW joint).

Figure 8
(a) 2D contour plot, (b) 3D contour plot (effect of TRS and W on UTS of FSW joint).

Figure 8 illustrates the interaction effect of W and TRS, with AL and WS held constant at 5 kN and 55 mm/min, respectively. The optimal UTS, around 494.81. MPa (as shown in both 2D and 3D plots), is achieved at a TRS of 1200 rpm and a W of 10 wt.% ZrO₂. Similarly, Figure 9 demonstrates the interaction effect of AL and WS, with TRS and W kept constant at 1200 rpm and 10 wt.%. The optimal UTS, approximately 494.81 MPa (evident in both 2D and 3D plots), occurs at a WS of 55 mm/min and an AL of 5 kN.

Figure 9
(a) 2D contour plot, (b) 3D contour plot (effect of AF and WS on UTS of FSW joint).

Figure 10 shows the interaction effect of W and WS, with AL and TRS held constant at 5 kN and 1200 rpm, respectively. The optimal UTS, around 494.81 MPa (as demonstrated in both 2D and 3D plots), is achieved at a WS of 55 mm/min and a W of 10 wt.% ZrO₂. Similarly, Figure 11 illustrates the interaction effect of W and AL, with TRS and WS kept constant at 1200 rpm and 55 mm/min, respectively. The optimal UTS, approximately 494.81 MPa (as shown in both 2D and 3D plots), is attained at a W of 10 wt.% ZrO₂ and an AL of 5 kN.

Figure 10
(a) 2D contour plot, (b) 3D contour plot (effect of W and WS on UTS of FSW joint).

Figure 11
(a) 2D contour plot, (b) 3D contour plot (effect of W and AF on UTS of FSW joint).

4.2. Weld Nugget Hardness (WNH)

The Hardness of the base AA6092 alloy, measured at 163 HRB, is lower than that of the stir zone. The weld nugget consistently exhibits higher Hardness than the base metal, irrespective of the tool's rotational speed. This increase in Hardness can be attributed to two main factors. First, the grain size in the stir zone is significantly finer than in the base metal, which strengthens the material according to the Hall-Petch equation, establishing a direct relationship between smaller grain size and increased Hardness. Second, the presence of fine intermetallic particles and uniformly dispersed ZrO₂ within the weld nugget further enhances its Hardness through Orowan hardening mechanisms. The difference in Hardness between the heat-affected zone (HAZ) and the stir zone is due to the specific grain refinement occurring in the stir zone.

Figure 12 (perturbation plot) illustrates the Effect of FSW process parameters on WNH for an optimized design. It is evident that across all operational levels of TRS, the weld nugget Hardness consistently exceeds that of the base metal. The plot also indicates that the Wt.% of ZrO₂ has the most significant influence on WNH, followed by TRS, AL, and WS.

Figure 12
Perturbation plot (effect of process parameters on the WNH).

The interaction effects of Tool Rotation Speed (TRS), Welding Speed (WS), Applied Load (AL), and Wt.% of ZrO₂ on Weld Nugget Hardness (WNH) are shown in Figures 13 through 18. Figures 13a and 13b illustrate the influence of TRS and WS on WNH while keeping AL constant at 5 kN. In the 2D contour plot (Figure 13a), the concentric circles represent Hardness (HRB), with the optimal value located at the centre of the plot. The maximum WNH of 184.46 HRB (Figures 13a and 13b) is achieved at a TRS of 1200 rpm and a WS of 55 mm/min. Figure 14 shows the interaction effect of AL and TRS at a constant WS of 55 mm/min and a W of 10 wt.% ZrO₂. The optimal WNH of approximately 184.46 HRB (as indicated in the 2D and 3D plots) occurs at a TRS of 1200 rpm and an AL of 5 kN. Figure 15 illustrates the interaction effect of W and WS, with a constant TRS of 1200 rpm and AL of 5 kN. The optimal WNH of 184.46 HRB (seen in both 2D and 3D plots) is achieved at a WS of 55 mm/min and a W of 10 wt.% ZrO₂.

Figure 13
(a) 2D contour plot, (b) 3D contour plot (effect of TRS and WS on WNH of FSW joint.

Figure 14
(a) 2D contour plot, (b) 3D contour plot (effect of TRS and AL on WNH of FSW joint).

Figure 15
(a) 2D contour plot, (b) 3D contour plot (effect of W and WS on WNH of FSW joint).

Figure 16 illustrates the effect of W and AL on the Weld Nugget Hardness (WNH) of the FSW joint, with a constant TRS of 1200 rpm and a WS of 55 mm/min. The optimal WNH, approximately 184.46 HRB (as shown in the 2D and 3D plots), is achieved at a W of 10 wt.% ZrO₂ and an AL of 5 kN. Figure 17 shows the influence of W and WS on WNH at a constant TRS of 1200 rpm and an AL of 5 kN. The optimal WNH, around 184.46 HRB (evident in the 2D and 3D plots), occurs at a WS of 55 mm/min and a W of 10 wt.% ZrO₂. Finally, Figure 18 highlights the impact of W and AL on WNH, with a constant TRS of 1200 rpm and a WS of 55 mm/min. The optimal WNH, approximately 184.46 HRB (visible in the 2D and 3D plots), is attained at an AL of 5 kN and a W of 10 wt.% ZrO₂.

Figure 16
(a) 2D contour plot, (b) 3D contour plot (effect of W and AL on WNH of FSW joint).

Figure 17
(a) 2D contour plot, (b) 3D contour plot (effect of W and WS on WNH of FSW joint).

Figure 18
(a) 2D contour plot, (b) 3D contour plot (impact of W and AL on WNH of FSW joint).

4.3. Percentage elongation

Figure 19 (perturbation plot) shows the influence of FSW process parameters on Percentage Elongation (PE) for an optimized design. Across all operational levels of TRS, the PE consistently exceeds that of the base metal. The plot clearly indicates that the weight percentage of ZrO₂ (W) is the most dominant factor affecting PE, followed by TRS, AL, and WS.

Figure 19
Perturbation plot (effect of process parameters on the PE).

The interaction effects of Tool Rotational Speed (TRS), Welding Speed (WS), Applied Load (AL), and weight percentage of ZrO₂ (W) on PE are depicted in Figures 20 through 25. Figures 20a and 20b illustrate the influence of TRS and WS on PE, with AL and W held constant at 5 kN and 10 wt.%, respectively. In the 2D contour plot (Figure 20a), the concentric circles represent PE (%), with the optimum value located at the centre of the plot. The maximum PE of 8.51% (Figures 20 a and 20 b) is achieved at a TRS of 1200 rpm and a WS of 55 mm/min.

Figure 20
(a) 2D contour plot, (b) 3D contour plot (effect of TRS and WS on PE of FSW joint).

Figure 21 depicts the interaction effect of AL and TRS, with WS kept constant at 55 mm/min and W at 10 wt.% ZrO₂. The optimal PE, approximately 8.51% (as seen in the 2D and 3D plots), is attained at a TRS of 1200 rpm and an AL of 5 kN. Figure 22 demonstrates the interaction effect of W and WS, with TRS held at 1200 rpm and AL at 5 kN. The optimal PE of 8.51% (evident in the 2D and 3D plots) is achieved at a WS of 55 mm/min and W of 10 wt.% ZrO₂.

Figure 21
(a) 2D contour plot, (b) 3D contour plot (effect of TRS and AL on PE of FSW joint).

Figure 22
(a) 2D contour plot, (b) 3D contour plot (effect of W and WS on PE of FSW joint).

Figure 23 illustrates the effect of W and AL on the Percentage Elongation (PE) of the FSW joint, with a constant TRS of 1200 rpm and a WS of 55 mm/min. The optimal PE value of approximately 8.51% (as shown in the 2D and 3D plots) is achieved at a W of 10 wt.% ZrO₂ and an AL of 5 kN.

Figure 23
(a) 2D contour plot, (b) 3D contour plot (effect of W and AL on PE of FSW joint).

Figure 24 shows the effect of W and WS on PE, with TRS held constant at 1200 rpm and AL at 5 kN. The optimal PE, around 8.51% (as depicted in the 2D and 3D plots), is observed at a WS of 55 mm/min and a W of 10 wt.% ZrO₂. Finally, Figure 25 highlights the effect of W and AL on PE at a constant TRS of 1200 rpm and a WS of 55 mm/min. The optimal PE value, approximately 8.51% (as indicated in the 2D and 3D plots), is attained at an AL of 5 kN and a W of 10 wt.% ZrO₂.

Figure 24
(a) 2D contour plot, (b) 3D contour plot (effect of W and WS on PE of FSW joint).

Figure 25
(a) 2D contour plot, (b) 3D contour plot (effect of W and AL on PE of FSW joint).

5. Results of Optimization

Following the optimization study aimed at achieving the desired mechanical properties for the welded joint, the optimum welding conditions were selected based on the optimization criteria, as outlined in Table 10. The experimental and optimization results suggest that a Tool Rotation Speed (TRS) of approximately 1200 rpm is ideal for achieving optimal Ultimate Tensile Strength (UTS), Percentage Elongation (PE), and Weld Nugget Hardness (WNH). Among the input parameters, the Wt.% of ZrO₂ has the most significant impact on these responses. The optimized FSW process parameters and the responses predicted by the design expert software are shown in Table 11. Figure 26 presents the contour and overlay plots, predicting the optimal UTS of 494.81 MPa, PE of 8.51%, and WNH of 184.57 HRB. These optimum values are achieved under the welding conditions of a TRS of 1208.09 rpm, a WS of 51.5 mm/min, and an AL of 5.13 kN.

Table 10
Optimization criteria used in this work.

Process Parameters / Responses	Goal	Minimum Limit	Maximum Limit	Importance
Tool Rotational Speed	is in range	1100	1300	3
Welding Speed	is in range	40	70	3
Axial Load	is in range	4	6	3
Wt.% ZrO2	is in range	5	15	3
Ultimate Tensile Strength	maximize	445	540	5
Percentage Elongation	maximize	4.98	9.56	5
Weld Nugget Hardness	maximize	151	195.5	5

Table 11
Optimized FSW process parameters and responses predicted by design expert software.

Tool rotational speed (rpm)	Welding Speed (mm/min)	Axial load (kN)	Wt.% of ZrO2	UTS (MPa)	PE (%)	Hardness (HRB)
1208.09	51.5	5.13	15	494.81	8.51	184.57

Figure 26
Contour plot and overlay plot for prediction plot of UTS, PE and WNH.

5.1. Validation of the developed model

The empirical or mathematical model developed using the desirability approach was validated by comparing its predictions with experimental results, and the errors were calculated for all thirty-one runs (Table 5). The experimental values were obtained through actual testing, while the predicted values were derived from the empirical equations generated by the design expert software. Table 12 presents the experimental values, predicted values, and percentage errors for UTS, PE, and weld nugget Hardness. For UTS, the percentage error ranges from +0.68% to +5.39%, for PE from -3.01% to +5.23%, and for WNH from -3.04% to +3.99%. These results demonstrate that the newly developed model accurately predicts UTS, PE, and WNH, as the predicted values closely match the experimental outcomes.

Table 12
Experimental and predicted values for thirty-one runs.

Run	UTS (MPa)			PE (%)			WNH (HRB)
	Experimental Value	Predicted Value	Error in %	Experimental Value	Predicted Value	Error in %	Experimental Value	Predicted Value	Error in %
1.	540.14	531.42	4.58	9.56	9.5	3.77	195.5	195.87	-0.19
2.	494.24	495.75	2.93	8.27	8.44	1.57	183	184.57	-0.86
3.	446.52	458.28	0.95	6.52	6.76	0.92	165.5	166.36	-0.52
4.	467.94	478.84	1.09	6.96	7.17	1.29	168	170.51	-1.49
5.	488.87	489.72	3.10	7.94	7.97	3.40	179	178.7	0.17
6.	474.05	464.48	5.39	7.52	7.73	1.20	175	168.02	3.99
7.	458.68	460.78	3.03	6.42	6.59	2.02	161.4	166.3	-3.04
8.	472.59	477.87	2.27	7.26	7.18	5.23	175	175.24	-0.14
9.	459.65	459.13	3.59	6.68	6.68	4.49	167	165.92	0.65
10.	472.32	468.97	4.10	7.09	7.48	-1.27	175	168.05	3.97
11.	493.26	495.75	2.74	8.22	8.44	0.97	183	184.57	-0.86
12.	495.45	495.75	3.17	8.42	8.44	3.33	185	184.57	0.23
13.	445.08	443.35	3.98	4.98	5.08	4.02	151	153.62	-1.74
14.	495.86	495.75	3.25	8.97	8.94	3.68	185	184.57	0.23
15.	453.64	451.53	3.99	5.85	6.09	1.03	158.5	156.75	1.10
16.	448.2	445.45	4.18	5.64	6.11	-3.01	158	155.27	1.73
17.	457.33	460.12	2.89	6.19	6.26	3.72	160.2	164.76	-2.85
18.	465.01	462.43	4.00	6.93	6.93	4.33	168	166.49	0.90
19.	489.18	491.49	2.80	8.01	8.11	2.50	179	180	-0.56
20.	493.26	495.75	2.74	8.22	8.44	0.97	183	184.57	-0.86
21.	461.21	457.65	4.24	6.9	6.97	3.33	167	163.95	1.83
22.	456.86	461.76	2.43	5.91	6.2	0.17	159	161.39	-1.50
23.	458.88	462.80	2.63	6.6	6.66	3.64	166	166.26	-0.16
24.	471.27	470.67	3.52	7.08	7.07	4.38	174	172.32	0.97
25.	495.45	495.75	3.17	8.42	8.44	3.33	184	184.57	-0.31
26.	467.64	480.46	0.68	6.96	7.24	0.29	170.1	174.34	-2.49
27.	482.35	489.85	1.76	7.61	7.76	1.97	177	178.89	-1.07
28.	474.13	466.07	5.07	7.54	7.88	-0.53	176	170.92	2.89
29.	486.06	487.38	3.02	7.62	7.69	3.02	178	177.47	0.30
30.	496.12	495.75	3.30	9.07	9.44	-0.77	189	184.57	2.34
31.	489.53	490.53	3.06	8.15	8.68	-2.82	179	175.77	1.80

The results of the validation experiments are summarized in Table 13. To assess the model's accuracy under the predicted optimal welding conditions, three confirmation experiments were carried out. These experiments employed a Tool Rotation Speed (TRS) of 1208.09 rpm, a Welding Speed (WS) of 51.5 mm/min, and an Applied Load (AL) of 5.13 kN. The maximum percentage errors observed in the predictions for Ultimate Tensile Strength (UTS), Percentage Elongation (PE), and Weld Nugget Hardness (WNH) were +0.04%, +0.47%, and +0.5%, respectively.

Table 13
Validation of test results.

Run	Tool Rotational Speed (rpm)	Welding Speed (mm/min)	Axial Load (kN)	UTS (MPa)		Error in %	Percentage Elongation (%)		Error in %	Weld Nugget Hardness (HRB)		Error in %
				Actual value	Predicted value		Actual value	Predicted value		Actual value	Predicted value
1	1208.09	51.5	5.13	494	494.81	-0.16	8.55	8.51	0.47	185	184.57	0.23
2	1208.09	51.5	5.13	495	494.81	0.04	8.54	8.51	0.35	184.5	184.57	-0.04
3	1208.09	51.5	5.13	493.5	494.81	-0.27	8.52	8.51	0.12	185.5	184.57	0.50

5.2. Microstructural characterization

Figure 27 a presents a macrostructural analysis of the friction stir (FS) welded joint. The cross-weld microstructure of the FS welded AA6092/ ZrO₂ joints, shown in Figures 27 b-d, reveals four distinct zones: the weld nugget zone (WNZ), the thermo-mechanically affected zone (TMAZ), the heat-affected zone (HAZ), and the unaffected zone (base material). This analysis highlights the absence of micron-level defects, attributed to sufficient heat generation and effective plastic flow during the welding process. Notably, the grains in the FS welded material are considerably finer than those in the base AA6092/ ZrO₂ Composite. The microstructure and grain size differ significantly across the WNZ, TMAZ, and HAZ due to varying heating and cooling conditions during the friction stir welding (FSW) process. Specifically, the WNZ displays finer grains compared to the TMAZ, HAZ, and unaffected zone. Additionally, the photomicrographs reveal fine recrystallized structures within the WNZ. The coarse grain structure observed in the base AA6092/ ZrO₂ Composite (Figure 27a) transitions to a finer grain structure as a result of the mechanical stirring action of the FSW tool, as illustrated in Figure 27d. This transformation in grain structure is a direct consequence of the FSW technique's influence on the microstructure of the welded material.

Figure 27
(a) Macrostructural analysis of FS welded composite and (b-d) Microstructural analysis of FS welded composite.

Figures 28 (a-d) present SEM micrographs of both the base AA6092/ ZrO₂ Composite and the FS welded alloy. Figure 28 (a) illustrates the grain size in the AA6092/ ZrO₂ Composite, while Figure 28b offers a clear view of the four distinct zones (TMAZ, WNZ, HAZ, and unaffected zone) within the FS welded composite. Figures 28c-d indicate that the WNZ has a finer grain structure compared to the TMAZ, HAZ, and unaffected zone. These micrographs emphasize the fine grain structure produced by the stirring action of the non-consumable rotating tool, which generates high plastic deformation and elevated temperatures. This process leads to dynamic recrystallization of grains, thereby enhancing the ultimate tensile strength (UTS) of the FS welded joints, although it may reduce ductility. The fine grain structures act as barriers to dendritic growth at the grain boundaries, resulting in improved UTS and Hardness of the base AA6092/ZrO₂ Composite. Furthermore, the micrographs demonstrate the accumulation and growth of clusters, which significantly enhance the overall strength of the AA6092/ ZrO₂ Composites. Grain size analysis revealed a decreasing trend with increasing ZrO₂ content, with the average grain size reducing from 21 µm (0 wt.%) to 12 µm (10 wt.%), attributed to particle-stimulated nucleation and Zener pinning effects.

Figure 28
(a-d) SEM analysis of AA6092/15% ZrO₂ composite and FS welded composite.

6. Conclusions

In this recent study, experiments were conducted to investigate various critical process parameters, including Tool Rotation Speed (TRS), Welding Speed (WS), Axial Load (AL), and weight percentage of ZrO₂, through multiple trials. The friction stir welding (FSW) process parameters for AA6092 composites were optimized, leading to several key findings:

❖ Regression models were developed using central composite design within Response Surface Methodology (RSM) to predict the Ultimate Tensile Strength (UTS), Percentage Elongation (PE), and Weld Nugget Hardness (WNH) of friction stir welded AA6092/ ZrO₂ composite joints.

❖ The accuracy of the developed regression models was validated through conformity tests.

❖ Perturbation plots, along with 2D and 3D contour plots, were analyzed using Design Expert software to assess the interaction effects of the welding parameters.

❖ Optimal values for UTS, PE, and WNH were achieved under specific welding conditions: a TRS of 1208.09 rpm, a WS of 51.5 mm/min, 15 wt.% ZrO₂, and an AL of 5.13 kN.

❖ The maximum UTS achieved (496.12 MPa) exceeds the values reported by Pandiyarajan et al. (2019), who observed ~470 MPa for AA6061/ZrO₂ composites. The increase can be attributed to better tool design, optimized FSW parameters, and uniform particle distribution in this study

❖ The maximum percentage errors for predicting optimal UTS, PE, and WNH were + 0.04%, + 0.47%, and + 0.5%, respectively.

• Data Availability
  The raw/processed data required to reproduce these findings cannot be shared at this time as the data also forms part of an ongoing study.

7. References

• ¹ Rajakumar S, Muralidharan C, Balasubramanian V. Optimization of the friction-stir-welding process and tool parameters to attain a maximum tensile strength of AA7075–T6 aluminum alloy. Proc Inst Mech Eng, B J Eng Manuf. 2010;224(8):1175-91. http://doi.org/10.1243/09544054JEM1802
  » http://doi.org/10.1243/09544054JEM1802
• ² Mishra RS, Ma ZY. Friction stir welding and processing. Mater Sci Eng Rep. 2005;50(1-2):1-78. http://doi.org/10.1016/j.mser.2005.07.001
  » http://doi.org/10.1016/j.mser.2005.07.001
• ³ Sharma C, Dwivedi DK, Kumar P. Effect of welding parameters on microstructure and mechanical properties of friction stir welded joints of AA7039 aluminum alloy. Mater Des. 2012;36:379-90. http://doi.org/10.1016/j.matdes.2011.10.054
  » http://doi.org/10.1016/j.matdes.2011.10.054
• ⁴ Abbass MK, Ameen HA, Hassan KS. Effect of heat treatment on corrosion resistance of friction stir welded AA 2024 aluminum alloy. Am J Sci Ind Res. 2011;2(2):297-306. http://doi.org/10.5251/ajsir.2011.2.2.297.306
  » http://doi.org/10.5251/ajsir.2011.2.2.297.306
• ⁵ Lakshminarayanan AK, Balasubramanian V, Elangovan K. Effect of welding processes on tensile properties of AA6061 aluminum alloy joints. Int J Adv Manuf Technol. 2009;40(3-4):286-96. http://doi.org/10.1007/s00170-007-1325-0
  » http://doi.org/10.1007/s00170-007-1325-0
• ⁶ Sayer S, Yeni C, Ertugrul O. Comparison of mechanical and microstructural behaviors of tungsten inert gas welded and friction stir welded dissimilar aluminum alloys AA 2014 and AA 5083. Kovove Mater. 2011;49(2):155-62. http://doi.org/10.4149/km_2011_2_155
  » http://doi.org/10.4149/km_2011_2_155
• ⁷ Fratini L, Buffa G, Filice L, Gagliardi F. Friction stir welding of AA6082-T6 T-joints: process engineering and performance measurement. Proc Inst Mech Eng, B J Eng Manuf. 2006;220(5):669-76. http://doi.org/10.1243/09544054JEM327
  » http://doi.org/10.1243/09544054JEM327
• ⁸ Scialpi A, Filippis LAC, Cavaliere P. Influence of shoulder geometry on microstructure and mechanical properties of friction stir welded 6082 aluminum alloy. Mater Des. 2007;28(4):1124-9. http://doi.org/10.1016/j.matdes.2006.01.031
  » http://doi.org/10.1016/j.matdes.2006.01.031
• ⁹ Vijayan D, Seshagiri Rao V. A parametric optimization of FSW process using RSM based grey relational analysis approach. Int Rev Mech Eng. 2007;8:328-37. http://doi.org/10.1007/s40032-014-0116-2.
• ¹⁰ Vijayan S, Raju R, Rao SRK. Mult objective optimization of friction stir welding process parameters on aluminum alloy AA 5083 using taguchi-based grey relation analysis. Mater Manuf Process. 2010;25(11):1206-12. http://doi.org/10.1080/10426910903536782
  » http://doi.org/10.1080/10426910903536782
• ¹¹ Kasman S. Multi-response optimization using the Taguchi-based grey relational analysis: a case study for dissimilar friction stir butt welding of AA6082-T6/AA5754-H111. Int J Adv Manuf Technol. 2013;68(1-4):795-804. http://doi.org/10.1007/s00170-012-4720-0
  » http://doi.org/10.1007/s00170-012-4720-0
• ¹² Prasanna P, Penchalayya Ch, Anandamohana Rao D. Optimization and validation of process parameters in friction stir welding on AA 6061 aluminum alloy using gray relational analysis. Int J Eng Res Appl. 2013;3:1471-81.
• ¹³ Shamsudeen S, Dhas JER. Optimization of multiple performance characteristics of friction stir welded joint with grey relational analysis. Mater Res. 2018;21(6):e20171050. http://doi.org/10.1590/1980-5373-mr-2017-1050
  » http://doi.org/10.1590/1980-5373-mr-2017-1050
• ¹⁴ El-Zathry NE, Hassan AI, El-Betar AA, Sabry I. Optimization of friction stir welding AA6082-T6 parameters using analysis of variance and grey relational analysis. J Phys Conf Ser. 2022;2299(1):012015. http://doi.org/10.1088/1742-6596/2299/1/012015
  » http://doi.org/10.1088/1742-6596/2299/1/012015
• ¹⁵ Dinaharan I, Murugan N. Influence of friction stir welding parameters on sliding wear behavior of AA6061/0-10 wt.% ZrB2 in-situ composite butt joints. J Miner Mater Charact Eng. 2011;10(14):1359-77. http://doi.org/10.4236/jmmce.2011.1014107
  » http://doi.org/10.4236/jmmce.2011.1014107
• ¹⁶ Palanivel R, Koshy Mathews P, Murugan N. Optimization of process parameters to maximize ultimate tensile strength of friction stir welded dissimilar aluminum alloys using response surface methodology. J Cent South Univ. 2013;20(11):2929-38. http://doi.org/10.1007/s11771-013-1815-1
  » http://doi.org/10.1007/s11771-013-1815-1
• ¹⁷ Ashok Kumar B, Murugan N. Optimization of friction stir welding process parameters to maximize tensile strength of stir cast AA6061-T6/AlN_p composite. Mater Des. 2014;57:383-93. http://doi.org/10.1016/j.matdes.2013.12.065
  » http://doi.org/10.1016/j.matdes.2013.12.065
• ¹⁸ Mohamed MA, Manurung Y, Berhan M. Model development for mechanical properties and weld quality class of friction stir welding using multi-objective Taguchi method and response surface methodology. J Mech Sci Technol. 2015;29(6):2323-31. http://doi.org/10.1007/s12206-015-0527-x
  » http://doi.org/10.1007/s12206-015-0527-x
• ¹⁹ Vijayan D, Rao VS. A parametric optimization of FSW process using RSM based grey relational analysis approach. Int Rev Mech Eng. 2016;8:328-37. http://doi.org/10.1007/s40032-014-0116-2.
• ²⁰ Faradonbeh AM, Shamanian M, Edris H, Paidar M, Bozkurt Y. Friction stir welding of Al-B4C composite fabricated by accumulative roll bonding: evaluation of microstructure and mechanical behavior. J Mater Eng Perform. 2017;27(2):835-46. http://doi.org/10.1007/s11665-018-3131-2
  » http://doi.org/10.1007/s11665-018-3131-2
• ²¹ Acharya U, Roy BS, Saha SC. A study of tool wear and its effect on the mechanical properties of friction stir welded AA6092/17.5 Sicp composite material joint. Mater Today Proc. 2018;5(9):20371-9. http://doi.org/10.1016/j.matpr.2018.06.412
  » http://doi.org/10.1016/j.matpr.2018.06.412
• ²² Pandiyarajan R, Maran P, Murugan N, Marimuthu S, Sornakumar T. Friction stir welding of hybrid AA 6061-ZrO2-C composites FSW process optimization using desirability approach’. Mater Res Express. 2019;6(6):1-16. http://doi.org/10.1088/2053-1591/ab084e
  » http://doi.org/10.1088/2053-1591/ab084e
• ²³ Salih OS, Ou H, Wei X, Sun W. Microstructure and mechanical properties of friction stir welded AA6092/SiC metal matrix composite. Mater Sci Eng A. 2019;742:78-88. http://doi.org/10.1016/j.msea.2018.10.116
  » http://doi.org/10.1016/j.msea.2018.10.116
• ²⁴ Singh T, Tiwari S, Shukla DK. Friction-stir welding of AA6061-T6: the effects of Al₂O₃ nanoparticles addition. Results in Materials. 2019;1:100005. http://doi.org/10.1016/j.rinma.2019.100005
  » http://doi.org/10.1016/j.rinma.2019.100005
• ²⁵ Singh T, Tiwari S, Shukla DK. Processing parameters optimization to produce nanocomposite using friction stir welding. Engineering Research Express. 2019;1(2):025015. http://doi.org/10.1088/2631-8695/ab5bee
  » http://doi.org/10.1088/2631-8695/ab5bee
• ²⁶ Singh T, Tiwari S, Shukla DK. Mechanical and microstructural characterization of friction stir welded AA6061-T6 joints reinforced with nano-sized particles. Mater Charact. 2019;159:110047. http://doi.org/10.1016/j.matchar.2019.110047
  » http://doi.org/10.1016/j.matchar.2019.110047
• ²⁷ Maneiah D, Mishra D, Rao KP, Raju KB. Process parameters optimization of friction stir welding for optimum tensile strength in Al 6061-T6 alloy butt welded joints. Mater Today Proc. 2020;27:904-8. http://doi.org/10.1016/j.matpr.2020.01.215
  » http://doi.org/10.1016/j.matpr.2020.01.215
• ²⁸ Singh T. Nanoparticles reinforced joints produced using friction stir welding: a review. Engineering Research Express. 2023;5(2):022001. http://doi.org/10.1088/2631-8695/accb28
  » http://doi.org/10.1088/2631-8695/accb28
• ²⁹ Singh T, Sharma Y. Microstructure evolution and mechanical strength mechanism in friction stir welded AA6061-T6/Al₂O₃ nanoparticles welds. Engineering Research Express. 2025;7(2):025003. http://doi.org/10.1088/2631-8695/adc3ec
  » http://doi.org/10.1088/2631-8695/adc3ec
• ³⁰ Umar Mohamed J, Palaniappan PLK, Maran P, Pandiyarajan R. Investigation and optimization of friction stir welding process parameters of stir cast AA6082/ ZrO₂ / B₄C composites. Mater Sci Pol. 2020;38(4):715-30. http://doi.org/10.2478/msp-2020-0082
  » http://doi.org/10.2478/msp-2020-0082
• ³¹ Umar Mohamed J, Palaniappan PLK, Maran P, Pandiyarajan R. Influences of ZrO₂ and B₄C reinforcement on metallurgical, mechanical, and tribological properties of AA6082 hybrid composite materials. J Ceram Process Res. 2021;22:306-16. http://doi.org/10.36410/jcpr.2021.22.3.306.
• ³² Jamaludeen UM. Examination and optimization of friction stir welding process parameters of AA6092 alloys. Soldag Insp. 2024;29:e2907. http://doi.org/10.1590/0104-9224/si29.07
  » http://doi.org/10.1590/0104-9224/si29.07
• ³³ Jamaludeen UM. Investigation and multi-response optimization of friction stir welding process parameters for stir cast AA6092/SiC composite. Mater Res. 2024;27:e20240228. http://doi.org/10.1590/1980-5373-mr-2024-0228
  » http://doi.org/10.1590/1980-5373-mr-2024-0228
• ³⁴ Jamaludeen UM. Development of empirical relationships for prediction of wear properties of AA6092/B₄ C aluminum matrix composites produced using friction stir welding. Mater Res. 2025;28:1-17.
• ³⁵ Jamaludeen UM. Analysis and multi-response optimization of friction stir welding parameters for stir-cast AA6092/B₄C/ZrO₂ hybrid composites. Mater Res. 2025;28:1-20. http://doi.org/10.1590/1980-5373-mr-2025-0185
  » http://doi.org/10.1590/1980-5373-mr-2025-0185

Edited by

• Associate Editor:
  José Daniel Biasoli de Mello.

• Editor-in-Chief:
  Luiz Antonio Pessan.

Data availability

The raw/processed data required to reproduce these findings cannot be shared at this time as the data also forms part of an ongoing study.

Publication Dates

• Publication in this collection
  05 Sept 2025
• Date of issue
  2025

History

• Received
  31 Mar 2025
• Reviewed
  24 June 2025
• Accepted
  16 July 2025