Microstructure based FE Simulation and Validation of Cold Upsetting on Aluminium Composites

Abstract

This study investigates the upset forging behavior of Aluminium 7068/SiC composite (AA 7068 with 5% vol. SiC) using the Finite Element Method (FEM). The effects of length-to-diameter ratio, friction coefficient, and initial relative density on the forging process are analyzed. FEM was applied to determine the bulge profile of the deformed billets during the upset forging process. Experiments were conducted to validate analytical models concerning the cold upset forging of solid cylindrical composite materials. Finite Element Method (FEM) simulations demonstrated strong correlation with the experimental data, exhibiting acceptable error margins for key deformation parameters. The study revealed significant dependencies of both hoop strain and axial stress on the specimen’s aspect ratio. Furthermore, an improvement in the formability stress index was observed with increasing axial strain. These findings offer valuable insights into the upset forging behavior of this class of composite materials.

Keywords:
AA 7068/SiC; Upsetting; FEM; Powder Metallurgy; Composite materials; ABAQUS

INTRODUCTION

The Finite Element Method (FEM) is a computer-based method of simulating and analyzing the behavior of engineering structures and components under a variety of conditions. It is an advanced engineering tool that is used in design and to replace experimental testing. Li et al. ¹ developed an upper bound solution for the determination of forging load and deformed bulged profile during upset forging of cylindrical billets, considering the dissimilar frictional conditions at flat die surfaces. Zhang et al. ² carried out an analysis of cold upsetting on cylindrical performs with aspect ratios. He found that the experimental values were in good agreement with the analytical results. Kucuk et al. ³ analyzed the plane strain upsetting process under two friction conditions using by boundary element method. Hu and Malayappan ^4)-(6 made an attempt to establish a relationship between the measured radius of curvature of the barrel and the new geometrical shape factor and hydrostatic stress during upset forging of solid cylinders with a die constraint at both ends. Manisekar and Narayanasamy ⁷ compared the effects of different lubricants while upsetting and found that molybdenum disulfide has effective lubrication compared with others tested. Syed Abu Thaheer and Narayanasamy ⁸ attempted to compare barreling in lubricated truncated cone billets. Ahasan et al. ⁹ made an analysis on the simulation and fracture prediction for sintered materials in upsetting by FEM. A finite element method program was developed to simulate the deformation processes of sintered materials in upsetting. The effects of lubrication, height to diameter ratio, initial relative density, die shaping, and preform shape on the forming limit of powder metallurgy products were investigated ¹⁰. AA7068 is a high-strength aluminum alloy belonging to the 7000 series, primarily composed of aluminum, zinc, magnesium, and copper. It is known for its exceptional mechanical properties, including high tensile strength, superior fatigue resistance, and good corrosion resistance, making it suitable for aerospace, automotive, and structural applications ^11)-(13. Silicon carbide (SiC) is a compound of silicon and carbon, known for its exceptional hardness, thermal stability, and electrical properties. It exists in various crystalline forms, with β-SiC (cubic) and α-SiC (hexagonal) being the most common. Due to its wide bandgap (3.26 eV) and high thermal conductivity (120-160 W/m·K), SiC is widely used in high-performance applications. Due to its hardness, SiC is widely used in grinding wheels, cutting tools, and wear-resistant coatings ^14)-(15. Murlidhar Patel ¹⁶ examined AA5052 metal matrix composites reinforced with 5 wt.% SiC particulates. The study reported a significant improvement in compressive strength and hardness compared to unreinforced AA5052, attributing the enhancement to the uniform dispersion of SiC particles. L. Pizzagalli and J. Godet explored the mechanical properties of SiC nanoparticles (1-2 nm) using first-principles molecular dynamics. The findings revealed that SiC nanoparticles exhibit ultrahigh strength, with compression simulations showing that theoretical bulk strength can be reached.¹⁷. Ameen Al Njjar ¹⁸ analyzed the effects of sintering temperature, compaction pressure, and SiC proportion on the density, hardness, and compressive strength of AA7075/SiC composites. The research found that 7.5% SiC reinforcement led to a 51.25% increase in compressive strength, making it highly desirable for aerospace applications. Powder Metallurgy (PM) is a kind of metal-forming process that involves creating components from metal powders. It’s a highly developed method for manufacturing both ferrous (iron-based) and non-ferrous materials, often used to produce parts with complex geometries and high precision. Powder metallurgy offers several significant advantages over traditional metal-forming techniques including Near net shape production, cost effectiveness, design flexibility and environment friendliness etc., yet, the drawback is its initial tooling and equipment costs ^19)-(21 While some studies integrate microstructural features like particle size, volume fraction into FE models, the direct and dynamic evolution of these microstructural features during cold upsetting and their real-time impact on macroscopic mechanical behavior was often simplified or not fully captured. Hence, in this study, an attempt has been made to analyze the behavior of Aluminium composites under upsetting using an FE model, and the flow characteristics have to be evaluated by experimental values.

This work aims to analyze the behavior of Aluminium composite (AA7068 with 5 % SiC) during upset forging using FEM. Validating the deformed profile generated by FEM using experimental data and studying the relationship between them.

MATERIALS AND METHODS

For FEM analysis, the commercially available software package ABAQUS is used. Only the first quadrant of the workpiece is calculated owing to symmetry. The working material was assumed to be solid, isotropic, and deformable. In the analysis, the workpiece is partitioned into two components as matrix and reinforcement, based on microstructure, as shown in Figure 1. Matrix is AA7068 with Young’s modulus of 70 GPa and Poisson’s ratio of 0.35, and reinforcement is SiC with Young’s modulus of 450 GPa and 0.17 Poisson’s ratio ^22)-(26. In the simulation, the matrix and reinforcement were modeled as distinct entities, each with independently defined material properties. A ‘tie’ constraint in ABAQUS was then applied to establish a monolithic behavior between them, and the temperature was set at room temperature throughout the analysis. The workpiece was considered fully dense, thereby excluding the influence of porosity or initial relative density on the simulated material behavior. A quadratic solid element mesh SOLID186 is utilized. The workpiece was finely meshed with 0.5 mm mesh and assembled with the dies. The FEM mesh shown is primarily quadrilateral, featuring high-quality elements in outer regions and concentrated, distorted elements near the regions where stress concentration is maximum. In calculating the power dissipation by friction at the die material interface, the friction factor should be determined. The ring compression test, a widely adopted method for determining friction factors in metal forming, was utilized. A friction factor of 0.72 was established for both dry and lubricated conditions, consistent with previously reported values. ^{7), (27}. The interaction between the workpiece and the die surface was defined by the contact elements. The bottom plate was kept fixed. The upper plate was set to move downwards at a speed of 0.5 mm/s, as upsetting is a gradual process. The time increment is 0.1s, and the increment of reduction is 1%.

Figure 1:
Microstructure of AA7068 with 5% SiC.

Aluminum alloy 7068 is a high-strength, heat-treatable alloy that offers excellent ductility and corrosion resistance. Its unique combination of properties makes it an ideal material for various applications. Silicon carbide (SiC) is a versatile ceramic material that offers a unique combination of properties, making it a valuable material in various industries. Materials for Aluminium alloy 7068 and SiC were purchased as powders from United Scientific Suppliers. The chemical composition utilized to make AA7068 alloy is provided in Table 1. Aluminum alloy 7068 with 5% volume SiC was blended thoroughly in a horizontal ball mill for 2 hours. The mixture is then compacted in a die set assembly and ejected as a solid cylinder of diameter 10 mm and 12 mm length. In order to prevent oxidation, the sample is coated with a mixture of Aluminium, alumina, and chalk powder and kept in open sunlight for 6 hrs. Finally, the sample is sintered at 500° C for 2 hours in a muffle furnace and allowed to furnace cool ^{28), (29}.

Table 1
Chemical Composition of AA7068

Element	Zn	Mg	Cu	Fe	Zr	Si	Mn	Ti	Cr	Al
Percentage	8.0	2.5	1.7	0.15	0.15	0.12	0.1	0.1	0.05	Rest

The prepared composite is then machined to have aspect ratios of 1, 0.75, and 0.5. Upset forging was carried out under dry conditions at room temperature using flat dies. Figure 2 shows the experimental setup for upset forging. This axial upsetting process was carried out using a Compression Testing Machine of 100 tons capacity. Much care was given to align the axis of the specimen with the axis of the die and platen. The load used during each deformation was recorded from the dial indicator of the Compression Testing Machine ^{30), (31}. The deformation process is stopped once a visible crack appears on the workpiece. Following each test, the height of the deformed specimen (h_f), bulge diameter (D_b), and contact diameters at the top and bottom (D_c1 and D_c2) were measured and recorded. Figure 3 displays the experimental samples at various stages, such as before heat treatment (a), before upsetting (b), and after upsetting (c).

Figure 2:
Experimental Setup.

Figure 3:
Samples of the experiment.

The following mathematical equations are considered under the triaxle stress state conditions. Based on the study of Rahman and El-Sheikh ³², the equations for calculating axial strain (ε_z) and the axial stress (σ_z) component of powder metallurgy composite preform can be calculated (Usually σ_z is negative because it is compressive) using the following expressions,

${\epsilon }_z=\ln \left(\frac{h_r}{h_o}\right)$ (1)

${\sigma }_z=\frac{Load}{Instantaneouscontactarea}$ (2)

according to Pereira et al. ³³, the von Mises stress under triaxial conditions can be given as:

${\sigma }_v=\sqrt{\frac{1}{2}\left[{\left({\sigma }_r-{\sigma }_{\theta }\right)}^2+[{\left({\sigma }_{\theta }-{\sigma }_z\right)}^2+[{\left({\sigma }_z-{\sigma }_r\right)}^2\right]+3\left({{\tau }^2}_{r\theta }+{{\tau }^2}_{\theta z}+{{\tau }^2}_{zr}\right)}$ (3)

where, s_r, s_q, s_z are radial normal stress, tangential normal stress and axial normal stresses respectively and t_rq, t_qz, t_zr are shear stresses acting along the planes in the specified directions

According to Raj et al. ³⁴, the true hoop strain (ε_θ) and hoop stress (σ_θ) can be expressed as follows,

${\epsilon }_{\theta }=\ln \left(\frac{2{D^2}_b+{D^2}_c}{2{D^2}_o}\right)$ (4)

${\sigma }_{\theta }=\left(\frac{2\alpha +R^2}{2-R^2+2\alpha R^2}\right){\sigma }_z$ (5)

The term α can be written as per Kumar et al. ³⁵ the state of stress as follow,

$\alpha =\frac{d{\epsilon }_{\theta }}{d{\epsilon }_z}=\frac{\left(2+R^2\right){\sigma }_{\theta }-R^2\left({\sigma }_z+2{\sigma }_{\theta }\right)}{\left(2+R^2\right){\sigma }_z-R^2\left({\sigma }_z+2{\sigma }_{\theta }\right)}$ (6)

When α, R, and σ_z are known, Eq. (5) can be used to calculate the hoop stress (σ_θ) can be rearranged as,

$\frac{{\sigma }_{\theta }}{{\sigma }_z}=\left(\frac{2\alpha +R^2}{2-R^2+2\alpha R^2}\right)$ (7)

Using the cylindrical coordinates (σ_θ = σ_z), the hydrostatic stress is calculated as,

${\sigma }_m=\frac{{\sigma }_r,{\sigma }_{\theta },{\sigma }_z}{3}=\frac{2{\sigma }_{\theta }+{\sigma }_z}{3},\text{and}\frac{{\sigma }_m}{{\sigma }_z}=\left(1+\frac{2{\sigma }_{\theta }}{{\sigma }_z}\right)$ (8)

Doraivelu et al. ³⁶ presented the equation of the effective stress as:

${\sigma }_1^2+{\sigma }_2^2+{\sigma }_3^2+-R^2\left({\sigma }_1{\sigma }_2+{\sigma }_2{\sigma }_3+{\sigma }_3{\sigma }_1\right)={\sigma }_e^2\left(2R^2-1\right)$ (9)

The above equation (eq. 9) can be written as for cylindrical axisymmetric upsetting in the cylindrical coordinates (σ_r = σ_θ),

${\sigma }_z^2+{\sigma }_{\theta }^2-R^2\left({\sigma }_{\theta }^2+2{\sigma }_{\theta }{\sigma }_z\right)={\sigma }_e^2\left(2R^2-1\right)$ (10)

Rearranging equation 10, the effective stress (σ_e) is given by

${\sigma }_e=\sqrt{\frac{{\sigma }_z^2+2{\sigma }_{\theta }^2-R^2\left({\sigma }_{\theta }^2+2{\sigma }_z{\sigma }_{\theta }\right)}{2R^2-1}}$ (11)

Vujovic and Shabaik ³⁷ proposed the equation for formability ratio (eq. 12), which gives the influence of hydrostatic stress and equivalent stress during the forging operation.

$\beta =3\left(\frac{{\sigma }_m/{\sigma }_z}{{\sigma }_e/{\sigma }_z}\right)$ (12)

Rearranging the equation 10, the formability stress index (β) is given as,

$\beta =\frac{3{\sigma }_m}{{\sigma }_e}$ (13)

According to Chiara Garavelli et al ³⁸, the formula to compare the experimental data with the FEM simulation results can be expressed as,

$\begin{array}{c}\text{Percentege}\\ \text{difference or error}\end{array}=\frac{\text{Experimental value}-\text{FEM value}}{\text{FEM value}}\times 100\%$ (14)

RESULTS AND DISCUSSIONS

The properties of Aluminium alloy AA7068 are significantly influenced by heat treatment, particularly tempering, and factors such as grain size, purity, and manufacturing process. A visual representation of the material behavior is depicted in Figure 4a, showcasing Silicon Carbide (SiC) particles dispersing outward, potentially illustrating phenomena like sputtering or crystal growth. For FEM analysis of AA 7068/SiC composite, a solid element mesh, preferably quadratic (SOLID186), is recommended to accurately capture stress gradients. The image presents a 2D FEM mesh with a predominantly quadrilateral structure, exhibiting a mix of high-quality elements in outer regions and distorted, clustered elements in areas of stress concentrations or geometric irregularities (Figure 4b).

Figure 4:
FEM Model and its Meshing.

The FEM analysis results for three different l/d ratios (1, 0.75, and 0.5) are presented, showcasing the von Mises stress distribution in MPa within a component (Figure 5). The color-coded representations reveal increasing stress magnitudes, with red indicating the highest stress concentrations. The meshes exhibit non-uniform structures, suggesting areas of varying stress concentration. The average stress levels are consistently indicated as 75%, providing a general measure of the state of the overall stress. The images effectively visualize the stress distribution, highlighting critical areas of potential failure or deformation within the analyzed component, with localized red regions corresponding to geometric discontinuities or points of load application.

Figure 5:
FEM Results for different (l/d) ratios.

During the experiment, it is observed that barreling takes place. This is due to friction between the workpiece and platens; the material faces a restraint in its flow at the top and bottom surfaces, whilst the middle portion flows freely. Figure 6 illustrates the relationship between axial strain and diametral strain for cylindrical specimens under compression, presented for three different length-to-diameter ratios (l/d): 1.0, 0.75, and 0.5. A linear trend is observed in all cases, indicating a direct correlation between axial and diametral strain, but the FEM simulations consistently underestimate the diametral strain compared to experimental measurements. In the case of l/d = 1, both curves show an increasing trend, indicating that as the axial strain increases, the true diametral strain also increases. This is expected behavior during deformation processes where material thinning is accompanied by lateral expansion. The experimental diametral strain values are consistently higher than the FEM diametral strain values across the entire range of h₀/h_f values, as shown. This suggests that the experimental setup or material behavior leads to greater diametral expansion than predicted by the FEM model. The percentage difference is not constant and tends to decrease as the axial strain increases, starting from over 40% at lower strains and reducing to around 25% at higher strains shown in the graph. This indicates that the FEM model’s prediction becomes somewhat closer to the experimental results at higher deformation levels, although a significant discrepancy still exists. For this instance, l/d = 0.75. At lower axial strains, the FEM values are higher than the experimental values. For instance, at $\ln \frac{h_0}{h_h}\approx 0.04$, the difference is approximately 33%, and at $\ln \frac{h_0}{h_f}\approx 0.08$, the difference is approximately 36%. This is a reversal of the trend observed in the l/d = 1, where experimental values were generally higher. As the true thickness strain increases, the FEM curve converges with the experimental curve. Notably, at $\ln \frac{h_0}{h_f}\approx 0.19$, there is excellent agreement, with a percentage difference of 0%. Beyond this point, at higher strains $\ln \frac{h_0}{h_f}\approx 0.25$, the FEM values are again slightly higher than the experimental values, with a difference of approximately 9%. In the case of l/d = 0.5 aspect ratio, the percentage difference starts around 10% at the lowest strain depicted, suggesting a relatively good agreement initially. However, as the axial strain increases, the discrepancy grows, reaching approximately 36% at $\ln \frac{h_0}{h_f}\approx 0.1$, before slightly decreasing to around 30% at the highest strain. This indicates that for an l/d ratio of 0.5, the FEM model deviates more significantly from experimental results at moderate thickness strains compared to very low or higher strains within the presented range. Based on the above analysis, it can be concluded that the discrepancy is most pronounced at l/d=1 and diminishes as the l/d ratio decreases, suggesting that specimen geometry significantly impacts the accuracy of the FEM model. This difference highlights potential limitations in the model’s material properties and underscores the importance of considering geometric effects in such analyses.

Figure 6:
Axial vs Diametral Strain for different (l/d) ratios.

Figure 7 depicts the relationship between axial strain and hoop strain for cylindrical specimens under compression, examined across three length-to-diameter ratios (l/d): 1.0, 0.75, and 0.5. Unlike a linear relationship, these graphs show a non-linear trend, with the curves flattening at higher axial strains, indicating plastic deformation. For an l/d ratio of 1, the experimental values of hoop strain are consistently higher than the FEM predicted values, especially at lower axial strains. Both curves show a non-linear increase in hoop strain with increasing axial strain. At lower strains, $\ln \frac{h_0}{h_f}\approx 0.02$, there’s a significant difference of approximately 40%, with FEM underpredicting the hoop strain. As the true thickness strain increases, the FEM predictions get progressively closer to the experimental values. The percentage difference reduces to about 20% at $\ln \frac{h_0}{h_f}\approx 0.07$, and further to less than 8% at higher strains. This suggests that the FEM model’s accuracy improves at higher deformation levels for this l/d ratio. Similar to l/d = 1, for an l/d ratio of 0.75, the experimental axial vs hoop strain values are generally higher than the FEM predictions. Both curves show a non-linear increasing trend. At lower strains $\ln \frac{h_0}{h_f}\approx 0.04$, the FEM model significantly underpredicts, with a difference of approximately 40%. As the axial strain increases, the agreement improves, with the percentage difference dropping to around 22% at $\ln \frac{h_0}{h_f}\approx 0.08$, and further reducing to about 14% at higher strains. This trend indicates that the FEM model’s prediction of hoop strain becomes more accurate at higher deformation levels for this aspect ratio as well. For an l/d ratio of 0.5, the experimental axial vs hoop strain values are generally higher than the FEM predictions, with both showing a non-linear increasing trend. The percentage difference starts around 10% at $\ln \frac{h_0}{h_f}\approx 0.035$. It then slightly increases to about 14% at $\ln \frac{h_0}{h_f}\approx 0.06$. It then decreases to around 12.5% at $\ln \frac{h_0}{h_f}\approx 0.095$. At the highest strain shown $\ln \frac{h_0}{h_f}\approx 0.13$, the difference is approximately 18%. Overall, the agreement between FEM and experimental results seems to be better for l/d = 0.5 compared to the other two aspect ratios, especially at the lower and mid-range of strains.

Figure 7:
Axial vs Hoop Strain for different (l/d) ratios.

Figure 8 illustrates the relationship between axial strain and various stresses (hoop, effective, axial, and hydrostatic) for cylindrical specimens under compression, examined across three length-to-diameter ratios (l/d): 1.0, 0.75, and 0.5. Each graph compares experimental data with FEM simulation results. Tensile stresses (effective and hoop) increase in magnitude with deformation, while compressive stresses (axial and hydrostatic) also increase in magnitude (become more negative) with deformation. This indicates that the FEM model successfully captures the qualitative behavior of stress evolution. The FEM predictions for effective stress consistently show very good agreement with experimental data across all l/d ratios. The percentage differences are generally less than 12%, and often well below 5%, especially for l/d = 0.5. This suggests that the chosen material model and overall simulation setup are quite effective in capturing the general flow behavior and plastic work. Similar to effective stress, hydrostatic stress also shows very good agreement between FEM and experimental results across all l/d ratios. The differences are typically less than 13%, often below 7%. This also indicates that the FEM model accurately predicts the volumetric stress state, which is crucial for predicting damage initiation. FEM consistently underpredicts experimental hoop stress. The discrepancies are more pronounced at lower deformation levels (e.g., up to 50% for l/d=1, 40% for l/d=0.75, and 25% for l/d=0.5). However, the agreement generally improves as the deformation increases, with differences reducing to less than 19% at higher strains for all l/d ratios. This suggests that while the overall material flow is captured, there might be subtle differences in the lateral expansion behavior that the FEM model struggles to perfectly replicate at initial stages. FEM generally overpredicts the magnitude, i.e., it predicts slightly less compressive axial stress compared to experimental values. The differences are usually within 20%, and often less than 10% at higher strains.

Figure 8:
Relationship between axial strain and different stresses

In summary, the aspect ratios 1.0 and 0.75 show slightly larger discrepancies for hoop and axial stresses, particularly at lower deformation levels, although the agreement generally improves with increasing strain. Whereas, for the aspect ratio 0.5, the agreement between FEM and experimental data for all stress components (especially effective and hydrostatic stress) is the best, with most percentage differences being within a tight range of 0-18%. This could imply that the chosen FEM setup or material model is particularly well-suited for specimens with this specific aspect ratio.

The comparison plots (Figure 9) reveal the formability stress index as it relates to axial strain for cylindrical specimens across varying aspect ratios (1.0, 0.75, 0.5), contrasting experimental and FEM simulation results. Both datasets consistently reveal that β remains relatively stable with increasing axial strain yet notably decreases as the aspect ratio lowers. This trend indicates a reduction in formability for shorter specimens. For an aspect ratio of 1.0, the experimental β values are consistently around 0.75, while FEM simulations predict values closer to 0.65. This seems that the FEM values are, on average, approximately 13.3% lower than the experimental values. For an aspect ratio of 0.75, experimental β values hover around 0.68, whereas FEM values are approximately 0.64. The FEM values are, on average, approximately 5.9% lower than the experimental values. For an aspect ratio of 0.5, the experimental β is around 0.57, while the FEM results are slightly lower at approximately 0.55. The resulting FEM values are approximately 3.5% lower than the experimental values. The polynomial fits for both experimental and FEM data, with high R² values ranging from 0.854 to 0.999 for experimental, and 0.871 to 0.921 for FEM, demonstrate a good correlation between Formability stress index β and axial strain within each dataset. However, the consistent discrepancy between the experimental and FEM values, particularly noticeable at higher aspect ratios, points to potential limitations in the FEM model.

Figure 9:
Comparison plots of Formability Stress Index (numerical and experimental results).

The contour plots (Figure 10) vividly compare the shapes and deformation profiles predicted by FEM and observed experimentally, demonstrating a sharp vertical drop in contour values across three distinct length-to-diameter (l/d) ratios: 1.0, 0.75, and 0.5. The x-axis represents the diameter, and the y-axis represents the height of the deformed specimen. Initially, for all aspects of ratios, the FEM and experimental lines align closely, indicating strong agreement in the material’s behavior under initial loading conditions. For instance, at l/d = 1.0, the initial height is approximately 7.7 units, maintained until a diameter of about 9.0 units. At l/d = 0.75, the initial height is approximately 5.6 units, maintained until a diameter of about 9.5 units. At l/d = 0.5, both FEM and experimental contours show an initial height of approximately 4.0 units and maintain this height until a diameter of about 9.5 units. Notably, the agreement between FEM and experimental contours generally improves as the l/d ratio decreases. At l/d = 1.0, the experimental curve deviates from the FEM curve around a diameter of 9.0 units, showing a slightly less abrupt or more gradual drop compared to FEM, which has a sharper, almost vertical drop around 9.5 units in diameter. At l/d = 0.75, the deviation is somewhat reduced, with both curves showing a sharp drop starting around 9.5 units in diameter, though the FEM curve appears slightly more vertical. The best match is observed at l/d = 0.5, where the FEM and experimental contours almost perfectly overlap throughout the deformation, including the sharp drop, which occurs consistently at a diameter of approximately 9.5 units. This validates the significant influence of specimen geometry (aspect ratio) on the accuracy of the FEM model’s predictions.

Figure 10:
Contour Plots FEM vs Experimental data.

CONCLUSIONS

The investigation successfully employed the FEM to simulate the cold upset forging behavior of AA 7068/SiC composite. The numerical results are validated with the help of experimental data to estimate the accuracy of the produced FEM model. The study revealed the effects of height-to-diameter ratio, friction coefficient, and initial relative density on the forging process. The findings of this research provide valuable insights into the cold upset forging behavior of AA 7068/SiC composite, highlighting the importance of FEM simulations in predicting material behavior and optimizing forging processes. FEM typically undervalues diametral strain, exhibiting errors between 10% and 41%, especially at elevated l/d ratios (1.0 and 0.5). Likewise, hoop strain was continuously underestimated by FEM, with substantial discrepancies of up to 40% at lower strain levels, which improved to 7-18% at elevated deformations. FEM exhibited substantial concordance with experimental data, with percentage discrepancies often ranging from 0-13%, indicating a reliable prediction of overall material flow and volumetric stress states. FEM frequently underestimated hoop stress, with discrepancies reaching 50% at low strains and diminishing to below 19% at high strains. Axial stress showed that FEM overestimated the amount by 3-20%. FEM consistently undervalued β across all l/d ratios. The disparity diminished with reduced l/d ratios, varying from roughly 13.3% for l/d=1.0 to 3.5% for l/d=0.5. This signifies enhanced model precision for shorter specimens. FEM offered an accurate qualitative depiction of the distorted geometries. The optimal accuracy in contour prediction was attained at l/d=0.5, demonstrating a strong correlation between FEM and experimental profiles. Deviations, especially at the abrupt decline signifying failure, become more evident at elevated l/d ratios.

ACKNOWLEDGEMENT

The authors would like to express their sincere gratitude to National Engineering College, Kovilpatti for providing the necessary resources and support to conduct this work.

DATA AVAILABILITY STATEMENT

Research data are only available upon request.

REFERENCES

• ¹ Li J, Zhou J, Li Y, Wei X. An analytical upper bound model for the prediction of load and deformed profile in multi-stage radial forging of tubes. Journal of Materials Processing Technology. 2020;275:116377. doi:10.1016/j.jmatprotec.2019.116377.
  » https://doi.org/10.1016/j.jmatprotec.2019.116377
• ² Zhang J, Dong F, Cao Y, Li M. Numerical and experimental investigation on cold upsetting of cylindrical billets with different aspect ratios. Journal of Manufacturing Processes. 2018;36:375-84. doi:10.1016/j.jmapro.2018.10.027.
  » https://doi.org/10.1016/j.jmapro.2018.10.027
• ³ Kucuk O, Cengiz Y, Gursel A. Numerical and experimental analysis of plane strain hot upsetting of AA6061 alloy. Journal of Mechanical Science and Technology. 2019;33(9):4257-65. doi:10.1007/s12206-019-0824-3.
  » https://doi.org/10.1007/s12206-019-0824-3
• ⁴ Hu Y, Yan Z, Jiang Z, Liu D, Li X. Influence of friction coefficient on the barreling behavior of Mg-Al-Zn alloy during hot compression. Materials Science and Engineering: A. 2022;850:143521. doi:10.1016/j.msea.2022.143521.
  » https://doi.org/10.1016/j.msea.2022.143521
• ⁵ Malayappan S, Narayanasamy R, Esakkimuthu G. Barrelling of aluminium solid cylinders during cold upset forging with constraint at both ends. Materials & Design. 2007;28(4):1404-11. doi:10.1016/j.matdes.2005.12.007.
  » https://doi.org/10.1016/j.matdes.2005.12.007
• ⁶ Jafari H, Shahriari D, Ebrahimi R. An investigation into friction and barreling in various open-die forging processes using the finite element method. Journal of Manufacturing Processes . 2018;32:337-48. doi:10.1016/j.jmapro.2018.03.011.
  » https://doi.org/10.1016/j.jmapro.2018.03.011
• ⁷ Manisekar K, Narayanasamy R. Effect of friction on barrelling in square and rectangular billets of aluminium during cold upset forging. Materials & Design . 2007;28(2):592-8. doi:10.1016/j.matdes.2005.07.017.
  » https://doi.org/10.1016/j.matdes.2005.07.017
• ⁸ Syed Abu Thaheer A, Narayanasamy R. Comparison of barreling in lubricated truncated cone billets during cold upset forging of various metals. Materials & Design . 2008;29:1027-35. doi:10.1016/j.matdes.2006.05.026.
  » https://doi.org/10.1016/j.matdes.2006.05.026
• ⁹ Ahasan M, Davidson M, Khanra AKK. Simulation studies on the densification behaviour of sintered aluminium preforms. Usak University Journal of Material Sciences. 2016;2(2):121-33. doi:10.12748/uujms.201324254.
  » https://doi.org/10.12748/uujms.201324254
• ¹⁰ Kannan AR, Pandey KS, Shanmugam S, Narayanasamy R, Arunachalam S, Page T. Deformation behaviour of Fe-0.8% C-1.0% Si-0.8% Cu sintered P/M steel during powder preform forging. i-Manager’s Journal on Future Engineering and Technology. 2010;5(3):54.
• ¹¹ Singh AK, Soni S, Rana RS, Kumar A, Singh RK, Chandra G, Srivastava SK. Tribological behavior of high-strength AA7068 alloy: effect of artificial aging temperature. NanoWorld Journal. 2023;12(1):14-22. doi:10.17756/nwj.2023-s1-100.
  » https://doi.org/10.17756/nwj.2023-s1-100
• ¹² Yazdani T, Soni A. Optimization of wear rate and coefficient of friction for AA7068/TiB₂/fly ash hybrid composite. SSRN. 2023. doi:10.2139/ssrn.4429997.
  » https://doi.org/10.2139/ssrn.4429997
• ¹³ Naganambi N, Selvakumar N. Investigation of mechanical and wear behaviour of AA7068 TiC graphene hybrid metal matrix composites. Journal of Materials Research and Technology. 2023;26:371-81. doi:10.1016/j.jmrt.2023.07.086.
  » https://doi.org/10.1016/j.jmrt.2023.07.086
• ¹⁴ La Via F, Alquier D, Giannazzo F, Kimoto T, Neudeck P, Ou H, Roncaglia A, Saddow SE, Tudisco S. Emerging SiC applications beyond power electronic devices. Materials. 2023;16(1):14. doi:10.3390/ma16010014.
  » https://doi.org/10.3390/ma16010014
• ¹⁵ Kimoto T, Yonezawa Y. Current status and perspectives of ultrahigh-voltage SiC power devices. Materials Science in Semiconductor Processing. 2018;78:43-56. doi:10.1016/j.mssp.2017.11.023.
  » https://doi.org/10.1016/j.mssp.2017.11.023
• ¹⁶ Patel M. SiC particulate reinforced AA5052 MMC. Journal of Modern Materials. 2021;12(3):45-52.
• ¹⁷ Pizzagalli L, Godet J. SiC nanoparticles and strength mechanisms. Physical Review Letters. 2014;112(1):015501. doi:10.1103/PhysRevLett.112.015501.
  » https://doi.org/10.1103/PhysRevLett.112.015501
• ¹⁸ Al Njjar A, Mazloum K, Sata A. Optimization of powder metallurgy parameters for improving the major properties of AA7075/SiC composites for aerospace applications. Journal of Materials Engineering and Performance. 2024;33:1-14. doi:10.1007/s11665-024-09998-z.
  » https://doi.org/10.1007/s11665-024-09998-z
• ¹⁹ Suryanarayanan S, Rao KP, Ravindran C, Sankaran S. Recent trends in powder metallurgy: a review. Materials Today: Proceedings. 2021;45:6019-24. doi:10.1016/j.matpr.2020.10.669.
  » https://doi.org/10.1016/j.matpr.2020.10.669
• ²⁰ Omar BZ, Abdullah B, Jamaludin SH. A comprehensive review on powder metallurgy process, its advantages, disadvantages, and applications. International Journal of Modern Manufacturing Technologies. 2021;13(2):52-63.
• ²¹ Singh RN, Puri D, Verma A. Recent developments in powder metallurgy of aluminum alloys: a review. Journal of Materials Research and Technology . 2019;8(5):4443-56. doi:10.1016/j.jmrt.2019.07.039.
  » https://doi.org/10.1016/j.jmrt.2019.07.039
• ²² Zhang P, Li F. Microstructure-based simulation of plastic deformation behavior of SiC particle reinforced Al matrix composites. Chinese Journal of Aeronautics. 2009;22(6):663-9. doi:10.1016/S1000-9361(08)60156-9.
  » https://doi.org/10.1016/S1000-9361(08)60156-9
• ²³ Fathipour M, Zoghipour P, Tarighi J, Yousefi R. Investigation of reinforced SiC particles percentage on machining force of metal matrix composite. Modern Applied Science. 2012;6(8):9-20. doi:10.5539/mas.v6n8p9.
  » https://doi.org/10.5539/mas.v6n8p9
• ²⁴ Prasanth Kumar B, Johoram Babu S. Microstructure based finite element analysis for deformation behaviour of aluminium based metal matrix composites. International Journal of Engineering Research & Technology. 2014;3:2221-3.
• ²⁵ Huang JY, Ling ZH, Wu GQ. Effects of particle-reinforcement on elastic modulus of metal-matrix composites. Materials Science Forum. 2010;650:285-9. doi:10.4028/www.scientific.net/MSF.650.285.
  » https://doi.org/10.4028/www.scientific.net/MSF.650.285
• ²⁶ Das M, Parimanik SR, Mahapatra TR, Mishra D. Effect of graphene reinforcement on the mechanical characteristics and machining performance of aluminium metal matrix composite. International Journal of Process Management and Benchmarking. 2023;14(4):478-507. doi:10.1504/IJPMB.2023.132216.
  » https://doi.org/10.1504/IJPMB.2023.132216
• ²⁷ Dehghan M, Qods F, Gerdooei M, Doai J. Analysis of ring compression test for determination of friction circumstances in forging process. Applied Mechanics and Materials. 2013;249:663-6. doi:10.4028/www.scientific.net/AMM.249-250.663.
  » https://doi.org/10.4028/www.scientific.net/AMM.249-250.663
• ²⁸ Alam M, Motgi BS. Study on microstructure and mechanical properties of Al7068 reinforced with silicon carbide and fly ash by powder metallurgy. International Journal for Modern Trends in Science and Technology. 2021;7(09):47-53. doi:10.46501/IJMTST0709009.
  » https://doi.org/10.46501/IJMTST0709009
• ²⁹ Joshua KJ, Sherjin P, Raj JP. Analysis of ball milled aluminium alloy 7068 metal powders. SSRG International Journal of Mechanical Engineering. 2017;4(8):6-10.
• ³⁰ Narayanasamy R, Ramesh T, Pandey KS. Some aspects on cold forging of aluminium-iron powder metallurgy composite under triaxial stress state condition. Materials & Design . 2008;29(4):891-903. doi:10.1016/j.matdes.2006.05.005.
  » https://doi.org/10.1016/j.matdes.2006.05.005
• ³¹ Hari Krishna C, Davidson MJ, Nagaraju C, Ramesh Kumar P. Effect of lubrication in cold upsetting using experimental and finite element modeling. Journal of Testing and Evaluation. 2015;43(1):53-61. doi:10.1520/JTE20130218.
  » https://doi.org/10.1520/JTE20130218
• ³² Rahman MA, El-Sheikh MN. Workability in forging of powder metallurgy compacts. Journal of Materials Processing Technology . 1995;54:97-102. doi:10.1016/0924-0136(95)01926-X.
  » https://doi.org/10.1016/0924-0136(95)01926-X
• ³³ Pereira K, Bordas S, Tomar S, Trobec R, Depolli M, Kosec G, Wahab MA. On the convergence of stresses in fretting fatigue. Materials. 2016;9(8):639. doi:10.3390/ma9080639.
  » https://doi.org/10.3390/ma9080639
• ³⁴ Raj AM, Selvakumar N, Narayanasamy R, Kailasanathan C. Experimental investigation on workability and strain hardening behaviour of Fe-C-Mn sintered composites with different percentage of carbon and manganese content. Materials & Design . 2013;49:791-801. doi:10.1016/j.matdes.2013.02.002.
  » https://doi.org/10.1016/j.matdes.2013.02.002
• ³⁵ Kumar DR, Narayanasamy R, Loganathan C. Effect of glass and SiC in aluminum matrix on workability and strain hardening behavior of powder metallurgy hybrid composites. Materials & Design . 2012;34:120-36. doi:10.1016/j.matdes.2011.07.062.
  » https://doi.org/10.1016/j.matdes.2011.07.062
• ³⁶ Doraivelu SM, Gegel HL, Gunasekera JS, Malas JC, Morgan JT, Thomas JF. A new yield function for compressible PM materials. International Journal of Mechanical Sciences. 1984;26(9-10):527-35. doi:10.1016/0020-7403(84)90006-7.
  » https://doi.org/10.1016/0020-7403(84)90006-7
• ³⁷ Vujovic V, Shabaik AH. A new workability criterion for ductile materials. Journal of Engineering Materials and Technology. 1986;108:245-9. doi:10.1115/1.3225876.
  » https://doi.org/10.1115/1.3225876
• ³⁸ Garavelli C, Curreli C, Palanca M, Aldieri A, Cristofolini L, Viceconti M. Experimental validation of a subject-specific finite element model of lumbar spine segment using digital image correlation. PLoS One. 2022;17(9):e0272529. doi:10.1371/journal.pone.0272529.
  » https://doi.org/10.1371/journal.pone.0272529

NOMENCLATURE

• ε_θ  Hoop strain
• ε_z  Axial strain
• σ_v  Von mises stress
• σ_θ  Hoop stress
• σ_z  Axial stress
• σ_r  Radial stress
• σ_m  Hydrostatic stress
• σ_e  Effective stress
• β  Formability ratio
• R  Relative density
• α  Poisson’s ratio
• dε_θ  Hoop strain increment during plastic deformation
• dε_z  Axial strain increment during plastic deformation
• h_o  Initial height
• h_f  Forged height
• D_b  Bulged diameter
• D_c  Contact diameter
• D_o  Initial diameter

Edited by

• (AE: Daniel Z. de Florio)

Publication Dates

• Publication in this collection
  17 Oct 2025
• Date of issue
  2025

History

• Received
  06 Mar 2025
• Reviewed
  25 June 2025
• Reviewed
  09 July 2025
• Accepted
  03 Aug 2025