Fig.3 Temperature fields at different steps:(a) 25th step; (b) 30th step
3.2 Temperature fields in thixoforging process
At the end of simulation of thixoforging process, five points are found in the part,as shown in Fig.4.

Fig.4 Schematic of intercept points
Fig.5 shows the curves of temperatures at different points. It is included that the temperature of the part decreases with the stage of thixoforging. The velocity of temperature decreases the fastest at Point P3. Points P1 and P2 were little higher than Point P3 because of plastic deformation. Points P3 and P4 were in the location of severe plastic deformation so the temperature at the beginning of thixoforging not only does not decline but also raises slowly, as shown in Fig.6 But when temperature reached a constant value, it begins to decrease. Because strain is high at the beginning of thixoforging, a large amount of heat is produced. When this point is out of the location of large deformation, the heat produced by deformation declines and the heat transfer by contacting begins to play a leading role.

Fig.5 Curves of temperature at different points

Fig.6 Drawing of temperature distribution
3.3 Effective strain fields in thixoforging process
Fig.7 shows the effective strain distribution fields. At the beginning of thixoforging process, severe deformation mostly appears around the corner of billet. A few of deformation occurs in the body of workpiece. At the second stage, the semi-solid billet pushes the sliding block forward with increasing the punch pressure. The deformation of Ф20 mm cylinder was bigger than that of Ф30 mm one. The both ends of the cylinder were severely deformed, and the whole billet had yielded to deformation. When the sliding block reached the maximum position, the effective strain was smaller than that at the last stage because of the hydraulic system.

Fig.7 Effective strain at different positions
Fig.8 shows the curves of effective strain at different points. It was concluded that the effective strain at the position which was contacted by punch was the smallest because of no plastic deformation, and it was regarded as rigid movement. The effective strains of points P1 and P2 rise slowly because of little plastic deformation. The effective strain of corner increased slowly at the beginning, and then increased rapidly, but at last decreased rapidly. It is because the two points of billet were in transfer area and plastic deformation was smaller at the beginning when they arrived at the corner, and high deformation was made. When deformation was in stable area, the plastic deformation declined, so the effective strain began to decrease.

Fig.8 Curves of effective strain at different points
3.4 Effective stress fields in thixoforging process
Fig.9 and Fig.10 show the effective stress distribution and the curves of effective stress at different points, respectively. It was known that the effective stress increased shortly at the beginning because it was in elastic deformation and the time was short. Elastic deformation was neglected and it was in accordance with the hypothesis. There was the large friction at Point P1 because of contact between Point P1 and punch, so the stress was the largest. Because the velocity of deformation was fast, the bigger friction was produced by workpiece and die. The stress of outside surface of workpiece was higher, and decreased gradually from outside to inside of workpiece. The stresses of Points P3 and P4 were much higher, which was caused by change of metal plastic deformation direction.
3.5 Velocity fields in thixoforging process
Fig.11 shows the velocity fields in the last stage. It was found that the velocity at the bottom of Ф50 mm cylinder was smaller than that at the two ends of part. The flowing of metal was not uniform, and the velocity of Ф30 mm cylinder was faster than that of Ф20 mm cylinder. It was in accordance with streamlines of workpiece, as shown in Fig.12.

Fig.9 Effective stresses at different positions

Fig.10 Curves of effective stress at different points

Fig.11 Velocity fields at last step

Fig.12 Streamline of workpiece
4 Experiments
4.1 Experimental procedure
In order to verify the regularity of plastic flowing, temperature and strain rate of thixoforging process of 6061 Al alloy, isothermal experiments were performed at 590 ?C for different holding times on 200 t hydraulic pressure machine. The hydraulic system shown in Fig.13 was introduced in this experiment as the assistance accomplishing the plastic deformation. The maximum limit pressure of overflow valve is designed to be 46 MPa and it was adjustable. Increasing the punch pressure to a certain value higher than overflow valve one, the sliding block was forced to move forward. Fig.14 shows the hydraulic stereogram.

Fig.13 Hydraulic schematic
4.2 Results of experiments and analysis
Fig.14 shows the workpiece from the experiment. The semi-solid billets made by SIMA were heated at 590 ?C for 50 min, then were put into the die and formed under the different pressures of side urn. The pressures of side urn were 46, 34 and 22 MPa, respectively.

Fig.14 Workpiece formed by semi-solid thixoforging

Fig.15 Microstructures of different points in Fig.4:(1) Point P1; (2) Point P2; (3) Point P3; (4) Point P4; (5) Point P5
Fig.15 shows the microstructures of different points as the same as those in Fig.4. The sizes of grains of Figs.15(c) and (d) are smaller than those in other images because of plastic deformation. The direction of plastic flowing can be seen from Figs.15(c) and (d), and the globular structure is filled with whole grain and distributed uniformly. These results are compatible with simulation. Table 1 shows the hardness of workpiece at different points formed by thixoforging. It is concluded that the hardness of Point P4 is the highest because of the severe plastic deformation, which is in accordance with simulation.
Table 1 Hardness comparison of workpiece at different points(HV)

5 Conclusions
1) The regularities of temperature, fluid and stress-strain fields are obtained.
2) The increasing pressure makes the sliding block move and the assistant hydraulic system put into effect, which leads to the real plastic deformation. The mechanical properties are improved largely by metallographic structure of workpiece. Both the strength and hardness are improved.
3) The globular structure is filled with whole grain and distributed uniformly. These results are compatible with simulation.
References
[1] EHRHEIM M, KOPP R. Automated process to produce a near-net -shape workpiece by open die forging [C]//Proceeding of the Fourth International Conference on Technology of Plasticity. Beijing, 1993: 1102?1107.
[2]DEAN T A. Progress in net-shape forging [C]// Proceeding of the Fourth International Conference on Technology of Plasticity. Beijing, 1993: 1031?1040.
[3]HIRAI K, SOMEKAWA H, TAKIGAWA Y, HIGASHIC K. Suprplastic forging with dynamic recrystallization of Mg-Al-Zn alloys cast by thixo-molding[J]. Scripta Materiala, 2007, 56: 237?240.
[4]PETROV P, PERFILOV V, STEBUNOV S. Prevention of lap formation in near net shape isothermal forging technology of part of irregular shape made of aluminium alloy A92618 [J]. Journal of Materials Processing Technology, 2006,177(1/2/3): 218?223.
[5]YOON J H, IM Y T, KIN N S. Finite element modeling of the deformation behavior of semi-solid materials[J]. Journal of Material Processing Technology, 2001, 113: 153?159.
[6]TOMEV B I, GAGOV V I. Modelling and description of the near-net-shape forging of cylindrical spur gears [J]. Journal of Materials Processing Technology, 1999, 92/93: 444?449.
[7] HUANG Guang-jie, ZHU Qing-Yan, HUANG Ben-duo. Research on constitutive relationship of 3104 aluminum alloy during plastic deformation in high temperature[J]. Materials Review, 2006, 20(s7): 460?462.
[8] LIU Lang-fei, PENG Da-shu, JIN Xiang-yun,et al.FEM simulation of extrusion of 3003 alloy tubes [J].Trans Nonferrous Met Soc China, 2000, 10(5): 656?659.
[9] WANG W W, SONG J L, HAN F, luo S J Numerical analysis on precision forging of 7A09 aluminum impeller [J].Material Science Forum, 2009, 628/629: 535?540.
[10] JI Ze-Sheng, YAN Hong-Hua, ZHENG Xiao-Ping, SUGIYAMA S O, YANAGIMOTO J. Numerical and experimental investigations of semi-solid AZ91D magnesium alloy in thixoforming process [J]. Journal of Materials Processing Technology, 2008, 8(5): 412?418.
[11] LI Xiao-li, LI Miao-qun, LI Feng, XIONG Ai-ming. Numerical simulation or grain size during isothermal forging of TC6 alloy [J]. The Chinese Journal of Nonferrous Metals, 2005, 15(9): 1332?1337.
(Edited by YANG Hua)
Foundation item: Projects(50875059, 50774026) supported by the National Natural Science Foundation of China; Project(20070420023) supported by China Postdoctoral Science Foundation; Project (2008AA03A239) supported by High-tech Research and Development Program of China
Corresponding author: DU Zhi-ming; Tel: +86-451-86415464; E-mail: duzm@263.net