A special extrusion-shear manufacturing method for magnesium alloy rods based on finite element numerical simulation and experimental verification

The influences of extrusion conditions and die structures on magnesium alloy rods prepared by extrusion-shear (ES) have been researched by finite element simulations and experiments. Computer finite element simulation software (DEFORM-3DTM finite element software) was used to simulate the evolution of load during the ES process with different preheated billet temperatures and strokes and channel angles. Microstructures of AZ31 magnesium alloy sampled from extruded rods prepared by direct extrusion and the ES process were observed to analyze deformation mechanisms. The results show that compared with the direct extrusion process, the ES process can significantly improve the equivalent strain value and deformation range of AZ31 magnesium alloy, thus refining the microstructure of magnesium alloy. Comparing the microstructure of the magnesium alloy at different preheating temperatures, it can be found that the grain growth rate of the ES process with a higher preheating temperature is significantly higher than that of the ES process.


Introduction
In the twenty-first century, magnesium alloys are considered green engineering materials. In recent years, magnesium alloys have been increasingly applied in many fields, such as the automotive, aerospace, and electronic industries. Additionally, the development opportunities of magnesium alloys will become more favorable over time [1]. Magnesium alloys are the lightest metal of all structural metallic materials, as their density is only 1.74 g/cm 3 , which is 2/3 that of aluminum alloys [2][3][4]. The characteristics of magnesium and magnesium alloys include low density, good damping, high-specific modulus, good heat conduction and electromagnetic shielding, and excellent cutting performance.
Magnesium alloys are deformed by rolling, forging, and extrusion [5][6][7][8]. Compared with rolling and forging, the size of the extruded products can be controlled in a quite accurate range, and the surface quality of the products is good.
Therefore, extrusion has been a significant way to produce magnesium alloy products. Although the traditional extrusion process tends to be mature, grain refinement of magnesium alloys is limited [9][10][11].
The material prepared by severe plastic deformation (SPD) technology has ultra-fine grained (UFG) structure, which can obtain UFG materials with grain sizes less than 1μm and form special texture, so it has excellent mechanical properties and performance. At present, the developed SPD technologies mainly include equal channel angular pressing (ECAP) [12,13], multi-directional forging (MDF) [14], high pressure torsion (HPT) [15,16], and reciprocating extrusion (CEC) [17]. Agnew et al. [18] studied the evolution of microstructure and properties of extruded AZ31B magnesium alloy after 8 passes of ECAP. The results show that the grain size of the alloy is refined from the initial 49 to 6μm. Sun et al. [19] studied the microstructure and property evolution of peak-aged Mg-8.2Gd-3.8Y-1.0Zn-0.4Zr alloy deformed by HPT at room temperature 6GPa. The results show that the grain size of the alloy is refined from 92μm in the initial state to 33nm in 16 cycles. Jiang et al. [20] studied the isothermal MDF deformation of as-cast AZ61 magnesium alloy at 300°C. The results showed that the grain size 1 3 of the alloy was refined from 320 to 3.7μm with the increasing deformation passes.
As a new technique for large plastic deformation, extrusion-shear (ES) is widely due to its simple mold structure and good forming effect. The extrusion force is the main factor that determines whether the billet can be successfully extruded during the ES process. The extrusion force is influenced by process parameters such as extrusion temperatures, extrusion speeds, extrusion ratios, and friction conditions for direct extrusion [21]. Moreover, it may also be affected by the alloy composition and die structures. This study aims to reveal the evolution of microstructures and explain the grain refinements mechanisms of AZ31 magnesium alloy during the ES process. The finite element software DEFORMTM-3D was used to simulate variations in extrusion loads during the ES process with different temperatures, strokes, and channel angles. The microstructures of AZ31 magnesium alloy obtained from rods prepared by direct extrusion and ES processes were observed.

Material property
The material used for simulations and experiments is commercial AZ31B magnesium alloy (Mg-3%Al-1%Zn, wt). The billets were annealed at 400°C for 15 h. Before extrusion, the billet is processed into a cylindrical blank with a diameter of 80 mm and a height of 50 mm. Before the experiment, the billet was placed in the extrusion cylinder and heated with the extrusion cylinder. After the temperature reached the specified temperature, extrusion was performed. After extrusion, the sample was cooled to room temperature by water cooling. Figure 1 is the schematic diagram of ES die structure, which is mainly composed of three parts: die, billet, and punch. The die structure, such as channel angle and extrusion ratio, plays an important role in the ES process. The three-dimensional extrusion model is shown in Fig. 2

Finite element simulation
The ES die realizes the deformation of the material by combining traditional extrusion and ECAP deformation. Compared with the traditional extrusion method, the deformation path, plastic flow behavior, and stress-strain state of the material are more complicated. Therefore, we use finite element numerical simulation software to analyze the metal flow law and deformation behavior of AZ31 magnesium alloy during ES process. The finite element analysis simulation was drawn by UG software and imported into DEFORM-3D TM software for simulation. To simplify the calculation process, the 1/2 model is used for simulation. In this paper, the mold uses the H13 hot-work die steel database in the software, and the billet uses AZ31 magnesium alloy. The AZ31 magnesium alloy material model is based on previous experimental results [22]. The blank is divided by tetrahedral mesh, the mesh ratio is 1.2, and the number of meshes is 80,000.
The Sparse solver and Newton-Paphson iteration method are used to solve the ES extrusion deformation process [23]. In the simulation process, the blank is a passive body and the punch is an active body. The extrusion temperatures were 350°C, 370°C, and 400°C, and the extrusion speed was 4mm/s. The remaining simulation parameters are shown in Table 2. Figure 2 shows the evolution of extrusion loads and equivalent strains during the ES process with preheated billet temperatures of 370°C. The ES process includes the upsetting stage, sizing stage, primary shearing stage, and secondary shearing stage. In the upsetting stage, the billet is put into the extrusion cylinder and in contact with the extrusion die. The outer layer of the billet begins to plastically deform first, and the outer layer of the billet that is first in contact with the die is pointed as shown in Fig. 2(a). As the extrusion continues, the value and range of the equivalent strain at the outer billet gradually diffuse, and the extrusion force also increases rapidly. After the billet enters the shear zone, the material shows uneven strain. The strain near the shear side is significantly higher than that on the inside, and the extrusion force also fluctuates. At the secondary shear corner, uniform strain along the billet interface can be observed. Finally, as the billet is extruded, the extrusion force decreases. Figure 3 shows the evolution of extrusion loads during the ES process with different preheated billet temperatures. At the upsetting stage of the ES process, extrusion loads are not affected by temperature rise [24]. Figure 3 shows that the extrusion loads of the ES process increase significantly with decreasing of preheated billet temperatures. The horizontal extruder with 300 tons of extrusion capacity in the laboratory can meet the extrusion forces required for the ES process with different  temperatures. During the primary and secondary shearing stages of the ES process with different preheated billet temperatures, the increase in extrusion loads is basically the same. Figure 4 shows the evolution curves of extrusion loads during the ES process with different channel angles. The extrusion load curves can be divided into four stages: upsetting, sizing, primary shearing, and secondary shearing. During the upsetting stage, the extrusion loads increase slowly. Plastic deformation of the billets has occurred. The loads reach maximum values, and the plastic deformation is not steady. As a result of dislocations accumulation, the loads increase rapidly due to work hardening. When the extrusion force slowly increases after 4.5 s, the extrusion curves are almost parallel to each other. As shown in the figures, compared with lager die channel angles, smaller die channel angles may lead to higher extrusion forces during the ES process.

Effects of extrusion methods on microstructures of magnesium alloy
The microstructures of the AZ31 magnesium alloy prepared by direct extrusion and ES processes are shown in Fig. 5. As shown in Fig. 5(a), there are grains with an average size of 50μm in the rods prepared by direct extrusion, which indicates that partial dynamic recrystallization (DRX) occurred during direct extrusion, and there are many original grains with grain sizes over 100μm. However, in Fig. 5(b), the grain refinement and homogenization of the magnesium alloy prepared by the ES process are very significant. Many tiny sub-grains appear around the original coarse grains, the average grain size was 12.15 μm. The DRXed fractions of the magnesium alloy prepared by the ES process are much higher than those prepared by direct extrusion. The accumulative strains increase and the grains are consequently during the ES process. It is clear that microstructures prepared by extrusion-shear are finer than those fabricated by direct extrusion. The microstructures of ES-extruded parts at different billet temperatures (370°C, 400°C, and 420°C) and channel angles of 135° are shown in Fig. 6(a), (b), and (c), and the parameters of the process and die structures for ES process are the same as those of reference [21]. As shown in Fig. 6(a), there is no uniform equiaxed grain in the ES hot extrusion rod. After the hot ES process, the grains are finer and more uniform. However, in Fig. 6(b) and (c), there are many original grains with sizes larger than 100μm. However, due to dynamic recrystallization (DRX), the fine grain size is approximately 10μm and the grain distribution is not uniform. It is obvious that the DRXed grain has grown.
The microstructures of the ES process with different billet temperatures of 370°C, 400°C, and 420°C are shown in Fig. 7, and the parameters of the process and die structures for ES process are the same as those of reference [21]. There are even equiaxed grains with an average grain size of approximately 10 μm in ES hot extruded rods, as shown in Fig. 7(a). This indicates that dense and sufficient DRX has taken place during the ES process when the billet preheating temperature is 370°C. After hot ES process, the grains are finer and more uniform. But in Fig. 7(b), there are many primordial particles with grain sizes greater than 100μm which indicates the occurrence of DRX. In addition, the grain size distribution is not uniform. In Fig. 7(d), due to the growth of DRXed grains, there are almost no fine DRXed particles in the billet. It is obvious that the grain growth rate of the ES process with a higher billet temperature is significantly higher than the grain refinement rate caused by the ES process [25].

Conclusions
In this paper, DEFORM-3D TM finite element software is used to simulate the ES process under different preheating billet temperatures and channel angles. The simulation results include the extrusion load, strain, and strain evolution in the ES process. The extrusion experiment and microstructure observation were carried out. The main results are summarized as follows: 1. Compared with the direct extrusion process, the ES process can significantly improve the equivalent strain value and deformation range of AZ31 magnesium alloy. The maximum strain value exceeds 5. Comparing the extrusion force of the ES process at different temperatures and different channel angles, the extrusion force of the ES process decreases with increasing billet preheating temperature and die channel angle. 2. By comparing the microstructure under different processes, it can be found that the ES process can introduce compressive strain and cumulative strain and effectively refine the microstructure of AZ31 magnesium alloy. After two shearing processes, the average grain size is reduced from 50 to 12.15 μm. 3. By comparing the microstructure of the ES process at different temperatures, it can be found that the grain growth rate of the ES process with a higher preheating temperature is significantly higher than the grain refinement rate caused by the ES process. Fig. 6 Microstructures of AZ31 Mg processed by the ES process with a channel angle of 135° and different preheated billet temperatures, a 370°C, b 400°C, and (c) 420°C, and the parameters are the same as those of reference [21] Fig. 7 Microstructures of AZ31 magnesium alloy processed by the ES process with a channel angle of 120° and different preheated billet temperatures: a 370°C; b 400°C, and (c) 420°C, and the parameters are the same as those of reference [21]