- Email: [email protected]
Effects of Key Simulation Parameters on Conical Ring Rolling Process
Available online at www.sciencedirect.com
ScienceDirect Procedia Engineering 81 (2014) 286 – 291
11th International Conference on Technology of Plasticity, ICTP 2014, 19-24 October 2014, Nagoya Congress Center, Nagoya, Japan
Effects of key simulation parameters on conical ring rolling process Wen Meng, Guo-qun Zhao* Key Laboratory for Liquid-Solid Structural Evolution and Processing of Materials (Ministry of Education), Shandong University, Jinan, Shandong 250061, China
Abstract This paper established 3D coupled thermo-mechanical finite element simulation model of a conical ring rolling process. Three simulation experiments were designed and performed. The effects of key simulation parameters such as mass scaling factor, time scaling factor and remeshing sweeps per increment on simulation time and volume change of conical ring were investigated and analyzed. The reasonable values of simulation parameters in conical ring rolling process were determined for decreasing simulation time and simultaneously reducing finite element simulation error. © 2014 2014 The Published by Elsevier Ltd. is anLtd. open access article under the CC BY-NC-ND license © Authors. Published by This Elsevier (http://creativecommons.org/licenses/by-nc-nd/3.0/). Selection and peer-review under responsibility of Nagoya University and Toyohashi University of Technology. Selection and peer-review under responsibility of the Department of Materials Science and Engineering, Nagoya University Keywords: Conical ring rolling; Simulation parameters; Simulation time; Volume change of conical ring
1. Introduction The radial conical ring rolling with a closed die structure on the top and bottom of driven roll is a branch of ring rolling. This rolling process is an advanced plastic forming technology that can produce seamless conical rings. It has advantages in saving materials and costs and improving production efficiency. The closed die structure on the top and bottom of driven roll can effectively prevent the generation of fishtail detects of conical rings and thus be frequently used in actual conical ring rolling processes by (Yuan, 2006) and (Han et al., 2007). The most typical conical products are nozzle supports, flanges (Seitz et al., 2013), aero-engine accessory, the main part of nuke reactor shell (Wang et al., 2007), and aluminum alloy conical ring with inner steps (Han et al., 2007). Recently, to decrease simulation time, many scholars simulated the ring rolling processes by changing the
* Corresponding author. Tel.: +86-531-88393238; fax: +86-531-88392811. E-mail address: [email protected]
1877-7058 © 2014 Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/3.0/). Selection and peer-review under responsibility of the Department of Materials Science and Engineering, Nagoya University doi:10.1016/j.proeng.2014.09.165
Wen Meng and Guo-qun Zhao / Procedia Engineering 81 (2014) 286 – 291
numbers of meshes, using advanced FE method, establishing appropriate optimization method and using mass scaling technology, respectively. Xu et al. (1991) set up a rigid plastic finite element analysis model of a ring rolling process, and analyzed the spreads and the pressure distribution in the roll gap. In order to reduce the simulation time, they selected different mesh numbers for the different analysis purposes. Lim et al. (1998) simulated rectangular and V-shaped ring rolling processes and improved computational time by 70% by using an implicit finite-strain updated-Lagrangian elastic-plastic FEM and a hybrid mesh model. Davey and Ward (2000) efficiently simulated ring rolling process by using arbitrary Lagrangian-Eulerian flow formulation and successive preconditioned conjugate gradient method. Wang et al. (2010) proposed an optimization objective function for the shortest simulation time, and their proposed optimization method reduced the simulation time by 28%. Qian et al. (2005) adopted mass scaling technology to decrease simulation time in rigid-plastic finite element simulation of ring rolling. The calculated results agreed well with simulation ones. To sum up, scholars mainly investigated the numerical simulation methods that can decrease simulation time in rectangular and V-shaped ring rolling processes. However, in the current studies, the effects of numerical simulation methods on the simulation accuracy were rarely discussed. And the studies about the effects of simulation parameters on conical ring rolling process were rarely reported. Therefore, it is necessary to investigate the effects of key simulation parameters such as mass scaling factor, time scaling factor and remeshing sweeps per increment on virtual conical ring rolling process. Based on ABAQUS software, this paper established the coupled thermo-mechanical 3D finite element model of a conical ring rolling process and investigated the effects of key simulation parameters on simulation time and volume change of the conical ring.
(a)
(b)
(c)
Fig. 1. (a) Dimensions of conical ring and rolls; (b) FE model of radial conical ring rolling; (c) simulation results.
2. FE modeling of conical ring rolling process The material Ti-6Al-4V is used in the FE model of the radial conical ring rolling process with a closed die structure on the top and bottom of driven roll. Its mechanical property and temperature-dependent physical properties are from reference (Yan et al., 2002). The conical ring is defined as a 3D deformable body; all rolls are defined as rigid bodies. Coupled thermo-displacement eight nodes six faces element is adopted to mesh the conical ring blank. The contact relationships between conical ring and rolls are defined. The frictional conditions and thermal conditions (thermal conduction, thermal convection and thermal radiation) are considered in the FE model. The friction factor between conical ring and rolls is 0.5. The temperature of initial conical blank is 900 °C. The
287
288
Wen Meng and Guo-qun Zhao / Procedia Engineering 81 (2014) 286 – 291
temperature of driven roll and guide rolls is 100 °C. The temperature of mandrel is 200 °C. The temperature of environment is 20 °C. The mass scaling technology, reduction integration technology and hourglass control technology are also employed. The feeding strategy that can realize a constant outer-radius-growth-rate of conical ring is adopted. The outer-radius-growth-rate of ring is selected as 4mm/s. The total feed amount of mandrel is 70 mm. Fig. 1(a) shows the dimensions of the ring and rolls. Fig. 1(b) shows the established simulation model for the conical ring rolling process with the closed die structure on the top and bottom of driven roll. Fig. 1(c) shows the simulation results the conical ring rolling process. 3. Design of numerical simulation experiments In numerical simulation of the conical ring rolling process, in order to achieve the convergence of the finite element solution, the time increment per step must be smaller than the critical time per step according to the reference presented by Qian et al. (2005).
't d 'tcr Cd
O
P
Lmin , Cd
O 2P , U Ev , (1 v)(1 2v) E , 2(1 v)
(1) (2) (3) (4)
where ¨t is the time increment per step; ¨t cr is the critical time per step; L min is the minimum of the distances between arbitrary two nodes of finite element meshes; C d is the speed of the stress wave propagation in finite element meshes; ȡ is mass destiny; Ȝ and ȝ are constants; E is Young’s modulus; v is Poisson’s ratio. Firstly, according to equations of (1) – (4), it can be deduced that when the mass density of material increases to m times, ¨t cr increases to m times. Accordingly, the total simulation time decreases to 1/ m times. However, a too larger mass scaling factor would result in a larger finite element simulation error. In this case, the simulation results cannot accurately reflect actual situation. Therefore, it is significantly necessary to determine an appropriate mass scaling factor. Secondly, the value of ¨t cr can also be changed by changing the value of time scaling factor in ABAQUS software. The default of the time scaling factor is 1. As the time scaling factor increases, ¨t cr is changed into a bigger value. In this case, the total simulation time decreases. It must be noted that when the time scaling factor is bigger than 1, the finite element solution may not be converged. Therefore, a reasonable value of the time scaling factor can reduce the simulation time. Finally, as the numerical simulation process proceeds, the finite element meshes are increasingly severely distorted. The finite element solution difficulties cased by severely distorted meshes will lead to the increase of simulation time. The numerical simulation can be even broken down by severely distorted meshes. However, the ALE remeshing technology can efficiently avoid these disadvantages. The number of remeshing per increment should be determined according to the degree of mesh distortion. The more severe the degree of mesh distortion is; the bigger the number of remeshing per increment is. In summary, according to above analysis, key simulation parameters (such as mass scaling factor, time scaling factor and remeshing sweeps per increment) have significant influences on both simulation time and simulation accuracy in the numerical simulation of the rolling process. In order to study the effects of key simulation parameters on the rolling process, the following three numerical simulation experiments were designed and performed. Case 1: In order to study the effects of mass scaling factor on the rolling process, experiment 1 simulated the rolling processes with five different values of mass scaling factor: (1) m = 50; (2) m = 100; (3) m = 200; (4) m = 300; (5) m = 400. The time scaling factor is selected as 1. The number of remeshing sweeps per increment is
Wen Meng and Guo-qun Zhao / Procedia Engineering 81 (2014) 286 – 291
selected as 1. Case 2: In order to study the effects of time scaling factor on the rolling process, experiment 2 simulated the rolling processes with five different values of time scaling factor: (1) n = 0.5; (2) n = 0.75; (3) n = 1.0; (4) n = 1.15; (5) n = 1.3. The mass scaling factor is selected as 100. The number of remeshing sweeps per increment is selected as 1. Case 3: In order to study the effects of remeshing sweeps per increment on the rolling process, experiment 3 simulated the rolling processes with four different values of remeshing sweeps per increment: (1) p = 1; (2) p = 5; (3) p = 10; (4) p = 15. The mass scaling factor is selected as 100. The time scaling factor is selected as 1. 4. Results and discussion Fig. 2 (a) shows the change curve of the simulation time over the mass scaling factor. The simulation time gradually decreases when the mass scaling factor increases from 50 to 200. This phenomenon agrees with calculation results in section 3 of this paper. However, the simulation time basically maintains constant when the mass scaling factor increases from 200 to 400. According to theoretical calculation, the simulation time should decrease. The reason of this phenomenon may be that when the mass scaling factor grows to 300 and 400, the finite element meshes are severely distorted. In this case, the finite element solution becomes more and more difficult. Therefore, the simulation time doesn’t decrease. Fig. 2 (b) shows the volume change curve of conical ring over the mass scaling factor. The volume loss of ring gradually increases as the mass scaling factor increases. This indicates that the finite element error grows up as the mass scaling factor increases. In conclusion, a too larger mass scaling factor (such as 300 and 400) is not able to save the simulation time and decrease the finite element error. An appropriate mass scaling factor for the established FE model of the rolling process is 50.
(a)
(b)
Fig. 2. Effects of mass scaling factor on conical ring rolling process. (a) Change curve of simulation time; (b) volume change curve of conical ring.
Fig. 3(a) shows the change curve of the simulation time over the time scaling factor. The simulation time gradually decreases as the time scaling factor increases. Fig. 3(b) shows the volume change curve of conical ring over the time scaling factor. The volume loss of conical ring firstly increases and then decreases, but the changing range of the volume loss of conical ring is small. When the time scaling factor equals to 1.3, the numerical simulation process was broken down because of non-convergence. This indicates that a larger time scaling factor can reduce simulation time, but a too larger time scaling factor may stop the numerical simulation process. An appropriate time scaling factor for the established FE model of the rolling process is 1.15.
289
290
Wen Meng and Guo-qun Zhao / Procedia Engineering 81 (2014) 286 – 291
(a)
(b)
Fig. 3. Effects of time scaling factor on conical ring rolling process. (a) Change curve of simulation time; (b) volume change curve of conical ring.
(a)
(b)
Fig. 4. Effects of remeshing sweeps per increment on conical rolling process. (a) Change curve of simulation time; (b) volume change curve of conical ring.
Fig. 4(a) shows the change curve of the simulation time over the remeshing sweeps per increment. The simulation time gradually increases as the number of remeshing per increment increases. Fig. 4(b) shows the volume change curve of the ring over the remeshing sweeps per increment. The volume loss of the ring is almost the same as the number of remeshing per increment increases. Therefore, an appropriate value of the number of remeshing per increment is 1. 5. Conclusions The effects of key simulation parameters (such as the mass scaling factor, the time scaling factor, and the remeshing sweeps per increment) on the volume loss of a conical ring and the simulation time of the conical ring rolling process were studied. In order to decrease simulation time and simultaneously reduce the finite element error of the simulation process of radial conical ring rolling process with a closed die structure on the top and bottom of driven roll, the appropriate values of the mass scaling factor, the time scaling factor and the number of remeshing per increment are 50, 1.15 and 1, respectively.
Wen Meng and Guo-qun Zhao / Procedia Engineering 81 (2014) 286 – 291
Acknowledgement This research work is supported by Program for Chang Jiang Scholars and Innovative Research Team in University of Ministry of Education of China (No. IRT0931). References Davey, K., Ward, M., 2000. An efficient solution method for finite element ring䇲㼞㼛㼘㼘㼕㼚㼓㻌 㼟㼕㼙㼡㼘㼍㼠㼕㼛㼚㻚㻌 㻵㼚㼠㼑㼞㼚㼍㼠㼕㼛㼚㼍㼘㻌 㻶㼛㼡㼞㼚㼍㼘㻌 㼒㼛㼞㻌 㻺㼡㼙㼑㼞㼕㼏㼍㼘㻌 Methods in Engineering 47, 1997-2018. Han, X.H., Hua, L., Lan, J., Zuo, Z.J., Jia, G.W., Huang, L.W., 2007. Simulation and experimental study of hot ring rolling of LD10 conical ring with inner steps. Journal of Wuhan University of Technology 29, 7-10. Lim, T., Pillinger, I., Hartley, P., 1998. A finite-element simulation of profile ring rolling using a hybrid mesh model. Journal of Materials Processing Technology 80, 199-205. Qian, D.S., Hua, L., Zuo, Z.J., Yuan, Y.L., 2005. Application of mass scaling in simulation of ring rolling by three-dimensional finite element method. Journal of Plasticity Engineering 12, 86-91. Seitz, J., Jenkouk, V., Hirt, G., 2013. Manufacturing dish shaped rings on radial-axial ring rolling mills. Production Engineering Research and Development 7, 611-618. Wang, Z.W., Fan, J.P., Hu, D.P., Tang, C.Y., Tsui, C.P., 2010. Complete modeling and parameter optimization for virtual ring rolling. International Journal of Mechanical Sciences 52, 1325-1333. Wang, Z.W., Zeng, S.Q., Yang, X.H., Cheng, C., 2007. The key technology and realization of virtual ring rolling. Journal of Materials Processing Technology 182, 374-381. Xu, S.G., Lian, J.C., Hawkyard, J.B., 1991. Simulation of ring rolling using a rigid-plastic finite element model. International Journal of Mechanical Sciences 33, 393-401. Yan, M.G., Liu, B.C., LI, J.G., 2002. China Aviation Material Manual. China Standard Press, Beijing,China. Yuan, H.L., 2006. The optimal design of profiled ring blank and rolling process numerical simulation. Dorctoral Dissertation for Huazhong University of Science and Technology Wuhan,China.
291
ScienceDirect Procedia Engineering 81 (2014) 286 – 291
11th International Conference on Technology of Plasticity, ICTP 2014, 19-24 October 2014, Nagoya Congress Center, Nagoya, Japan
Effects of key simulation parameters on conical ring rolling process Wen Meng, Guo-qun Zhao* Key Laboratory for Liquid-Solid Structural Evolution and Processing of Materials (Ministry of Education), Shandong University, Jinan, Shandong 250061, China
Abstract This paper established 3D coupled thermo-mechanical finite element simulation model of a conical ring rolling process. Three simulation experiments were designed and performed. The effects of key simulation parameters such as mass scaling factor, time scaling factor and remeshing sweeps per increment on simulation time and volume change of conical ring were investigated and analyzed. The reasonable values of simulation parameters in conical ring rolling process were determined for decreasing simulation time and simultaneously reducing finite element simulation error. © 2014 2014 The Published by Elsevier Ltd. is anLtd. open access article under the CC BY-NC-ND license © Authors. Published by This Elsevier (http://creativecommons.org/licenses/by-nc-nd/3.0/). Selection and peer-review under responsibility of Nagoya University and Toyohashi University of Technology. Selection and peer-review under responsibility of the Department of Materials Science and Engineering, Nagoya University Keywords: Conical ring rolling; Simulation parameters; Simulation time; Volume change of conical ring
1. Introduction The radial conical ring rolling with a closed die structure on the top and bottom of driven roll is a branch of ring rolling. This rolling process is an advanced plastic forming technology that can produce seamless conical rings. It has advantages in saving materials and costs and improving production efficiency. The closed die structure on the top and bottom of driven roll can effectively prevent the generation of fishtail detects of conical rings and thus be frequently used in actual conical ring rolling processes by (Yuan, 2006) and (Han et al., 2007). The most typical conical products are nozzle supports, flanges (Seitz et al., 2013), aero-engine accessory, the main part of nuke reactor shell (Wang et al., 2007), and aluminum alloy conical ring with inner steps (Han et al., 2007). Recently, to decrease simulation time, many scholars simulated the ring rolling processes by changing the
* Corresponding author. Tel.: +86-531-88393238; fax: +86-531-88392811. E-mail address: [email protected]
1877-7058 © 2014 Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/3.0/). Selection and peer-review under responsibility of the Department of Materials Science and Engineering, Nagoya University doi:10.1016/j.proeng.2014.09.165
Wen Meng and Guo-qun Zhao / Procedia Engineering 81 (2014) 286 – 291
numbers of meshes, using advanced FE method, establishing appropriate optimization method and using mass scaling technology, respectively. Xu et al. (1991) set up a rigid plastic finite element analysis model of a ring rolling process, and analyzed the spreads and the pressure distribution in the roll gap. In order to reduce the simulation time, they selected different mesh numbers for the different analysis purposes. Lim et al. (1998) simulated rectangular and V-shaped ring rolling processes and improved computational time by 70% by using an implicit finite-strain updated-Lagrangian elastic-plastic FEM and a hybrid mesh model. Davey and Ward (2000) efficiently simulated ring rolling process by using arbitrary Lagrangian-Eulerian flow formulation and successive preconditioned conjugate gradient method. Wang et al. (2010) proposed an optimization objective function for the shortest simulation time, and their proposed optimization method reduced the simulation time by 28%. Qian et al. (2005) adopted mass scaling technology to decrease simulation time in rigid-plastic finite element simulation of ring rolling. The calculated results agreed well with simulation ones. To sum up, scholars mainly investigated the numerical simulation methods that can decrease simulation time in rectangular and V-shaped ring rolling processes. However, in the current studies, the effects of numerical simulation methods on the simulation accuracy were rarely discussed. And the studies about the effects of simulation parameters on conical ring rolling process were rarely reported. Therefore, it is necessary to investigate the effects of key simulation parameters such as mass scaling factor, time scaling factor and remeshing sweeps per increment on virtual conical ring rolling process. Based on ABAQUS software, this paper established the coupled thermo-mechanical 3D finite element model of a conical ring rolling process and investigated the effects of key simulation parameters on simulation time and volume change of the conical ring.
(a)
(b)
(c)
Fig. 1. (a) Dimensions of conical ring and rolls; (b) FE model of radial conical ring rolling; (c) simulation results.
2. FE modeling of conical ring rolling process The material Ti-6Al-4V is used in the FE model of the radial conical ring rolling process with a closed die structure on the top and bottom of driven roll. Its mechanical property and temperature-dependent physical properties are from reference (Yan et al., 2002). The conical ring is defined as a 3D deformable body; all rolls are defined as rigid bodies. Coupled thermo-displacement eight nodes six faces element is adopted to mesh the conical ring blank. The contact relationships between conical ring and rolls are defined. The frictional conditions and thermal conditions (thermal conduction, thermal convection and thermal radiation) are considered in the FE model. The friction factor between conical ring and rolls is 0.5. The temperature of initial conical blank is 900 °C. The
287
288
Wen Meng and Guo-qun Zhao / Procedia Engineering 81 (2014) 286 – 291
temperature of driven roll and guide rolls is 100 °C. The temperature of mandrel is 200 °C. The temperature of environment is 20 °C. The mass scaling technology, reduction integration technology and hourglass control technology are also employed. The feeding strategy that can realize a constant outer-radius-growth-rate of conical ring is adopted. The outer-radius-growth-rate of ring is selected as 4mm/s. The total feed amount of mandrel is 70 mm. Fig. 1(a) shows the dimensions of the ring and rolls. Fig. 1(b) shows the established simulation model for the conical ring rolling process with the closed die structure on the top and bottom of driven roll. Fig. 1(c) shows the simulation results the conical ring rolling process. 3. Design of numerical simulation experiments In numerical simulation of the conical ring rolling process, in order to achieve the convergence of the finite element solution, the time increment per step must be smaller than the critical time per step according to the reference presented by Qian et al. (2005).
't d 'tcr Cd
O
P
Lmin , Cd
O 2P , U Ev , (1 v)(1 2v) E , 2(1 v)
(1) (2) (3) (4)
where ¨t is the time increment per step; ¨t cr is the critical time per step; L min is the minimum of the distances between arbitrary two nodes of finite element meshes; C d is the speed of the stress wave propagation in finite element meshes; ȡ is mass destiny; Ȝ and ȝ are constants; E is Young’s modulus; v is Poisson’s ratio. Firstly, according to equations of (1) – (4), it can be deduced that when the mass density of material increases to m times, ¨t cr increases to m times. Accordingly, the total simulation time decreases to 1/ m times. However, a too larger mass scaling factor would result in a larger finite element simulation error. In this case, the simulation results cannot accurately reflect actual situation. Therefore, it is significantly necessary to determine an appropriate mass scaling factor. Secondly, the value of ¨t cr can also be changed by changing the value of time scaling factor in ABAQUS software. The default of the time scaling factor is 1. As the time scaling factor increases, ¨t cr is changed into a bigger value. In this case, the total simulation time decreases. It must be noted that when the time scaling factor is bigger than 1, the finite element solution may not be converged. Therefore, a reasonable value of the time scaling factor can reduce the simulation time. Finally, as the numerical simulation process proceeds, the finite element meshes are increasingly severely distorted. The finite element solution difficulties cased by severely distorted meshes will lead to the increase of simulation time. The numerical simulation can be even broken down by severely distorted meshes. However, the ALE remeshing technology can efficiently avoid these disadvantages. The number of remeshing per increment should be determined according to the degree of mesh distortion. The more severe the degree of mesh distortion is; the bigger the number of remeshing per increment is. In summary, according to above analysis, key simulation parameters (such as mass scaling factor, time scaling factor and remeshing sweeps per increment) have significant influences on both simulation time and simulation accuracy in the numerical simulation of the rolling process. In order to study the effects of key simulation parameters on the rolling process, the following three numerical simulation experiments were designed and performed. Case 1: In order to study the effects of mass scaling factor on the rolling process, experiment 1 simulated the rolling processes with five different values of mass scaling factor: (1) m = 50; (2) m = 100; (3) m = 200; (4) m = 300; (5) m = 400. The time scaling factor is selected as 1. The number of remeshing sweeps per increment is
Wen Meng and Guo-qun Zhao / Procedia Engineering 81 (2014) 286 – 291
selected as 1. Case 2: In order to study the effects of time scaling factor on the rolling process, experiment 2 simulated the rolling processes with five different values of time scaling factor: (1) n = 0.5; (2) n = 0.75; (3) n = 1.0; (4) n = 1.15; (5) n = 1.3. The mass scaling factor is selected as 100. The number of remeshing sweeps per increment is selected as 1. Case 3: In order to study the effects of remeshing sweeps per increment on the rolling process, experiment 3 simulated the rolling processes with four different values of remeshing sweeps per increment: (1) p = 1; (2) p = 5; (3) p = 10; (4) p = 15. The mass scaling factor is selected as 100. The time scaling factor is selected as 1. 4. Results and discussion Fig. 2 (a) shows the change curve of the simulation time over the mass scaling factor. The simulation time gradually decreases when the mass scaling factor increases from 50 to 200. This phenomenon agrees with calculation results in section 3 of this paper. However, the simulation time basically maintains constant when the mass scaling factor increases from 200 to 400. According to theoretical calculation, the simulation time should decrease. The reason of this phenomenon may be that when the mass scaling factor grows to 300 and 400, the finite element meshes are severely distorted. In this case, the finite element solution becomes more and more difficult. Therefore, the simulation time doesn’t decrease. Fig. 2 (b) shows the volume change curve of conical ring over the mass scaling factor. The volume loss of ring gradually increases as the mass scaling factor increases. This indicates that the finite element error grows up as the mass scaling factor increases. In conclusion, a too larger mass scaling factor (such as 300 and 400) is not able to save the simulation time and decrease the finite element error. An appropriate mass scaling factor for the established FE model of the rolling process is 50.
(a)
(b)
Fig. 2. Effects of mass scaling factor on conical ring rolling process. (a) Change curve of simulation time; (b) volume change curve of conical ring.
Fig. 3(a) shows the change curve of the simulation time over the time scaling factor. The simulation time gradually decreases as the time scaling factor increases. Fig. 3(b) shows the volume change curve of conical ring over the time scaling factor. The volume loss of conical ring firstly increases and then decreases, but the changing range of the volume loss of conical ring is small. When the time scaling factor equals to 1.3, the numerical simulation process was broken down because of non-convergence. This indicates that a larger time scaling factor can reduce simulation time, but a too larger time scaling factor may stop the numerical simulation process. An appropriate time scaling factor for the established FE model of the rolling process is 1.15.
289
290
Wen Meng and Guo-qun Zhao / Procedia Engineering 81 (2014) 286 – 291
(a)
(b)
Fig. 3. Effects of time scaling factor on conical ring rolling process. (a) Change curve of simulation time; (b) volume change curve of conical ring.
(a)
(b)
Fig. 4. Effects of remeshing sweeps per increment on conical rolling process. (a) Change curve of simulation time; (b) volume change curve of conical ring.
Fig. 4(a) shows the change curve of the simulation time over the remeshing sweeps per increment. The simulation time gradually increases as the number of remeshing per increment increases. Fig. 4(b) shows the volume change curve of the ring over the remeshing sweeps per increment. The volume loss of the ring is almost the same as the number of remeshing per increment increases. Therefore, an appropriate value of the number of remeshing per increment is 1. 5. Conclusions The effects of key simulation parameters (such as the mass scaling factor, the time scaling factor, and the remeshing sweeps per increment) on the volume loss of a conical ring and the simulation time of the conical ring rolling process were studied. In order to decrease simulation time and simultaneously reduce the finite element error of the simulation process of radial conical ring rolling process with a closed die structure on the top and bottom of driven roll, the appropriate values of the mass scaling factor, the time scaling factor and the number of remeshing per increment are 50, 1.15 and 1, respectively.
Wen Meng and Guo-qun Zhao / Procedia Engineering 81 (2014) 286 – 291
Acknowledgement This research work is supported by Program for Chang Jiang Scholars and Innovative Research Team in University of Ministry of Education of China (No. IRT0931). References Davey, K., Ward, M., 2000. An efficient solution method for finite element ring䇲㼞㼛㼘㼘㼕㼚㼓㻌 㼟㼕㼙㼡㼘㼍㼠㼕㼛㼚㻚㻌 㻵㼚㼠㼑㼞㼚㼍㼠㼕㼛㼚㼍㼘㻌 㻶㼛㼡㼞㼚㼍㼘㻌 㼒㼛㼞㻌 㻺㼡㼙㼑㼞㼕㼏㼍㼘㻌 Methods in Engineering 47, 1997-2018. Han, X.H., Hua, L., Lan, J., Zuo, Z.J., Jia, G.W., Huang, L.W., 2007. Simulation and experimental study of hot ring rolling of LD10 conical ring with inner steps. Journal of Wuhan University of Technology 29, 7-10. Lim, T., Pillinger, I., Hartley, P., 1998. A finite-element simulation of profile ring rolling using a hybrid mesh model. Journal of Materials Processing Technology 80, 199-205. Qian, D.S., Hua, L., Zuo, Z.J., Yuan, Y.L., 2005. Application of mass scaling in simulation of ring rolling by three-dimensional finite element method. Journal of Plasticity Engineering 12, 86-91. Seitz, J., Jenkouk, V., Hirt, G., 2013. Manufacturing dish shaped rings on radial-axial ring rolling mills. Production Engineering Research and Development 7, 611-618. Wang, Z.W., Fan, J.P., Hu, D.P., Tang, C.Y., Tsui, C.P., 2010. Complete modeling and parameter optimization for virtual ring rolling. International Journal of Mechanical Sciences 52, 1325-1333. Wang, Z.W., Zeng, S.Q., Yang, X.H., Cheng, C., 2007. The key technology and realization of virtual ring rolling. Journal of Materials Processing Technology 182, 374-381. Xu, S.G., Lian, J.C., Hawkyard, J.B., 1991. Simulation of ring rolling using a rigid-plastic finite element model. International Journal of Mechanical Sciences 33, 393-401. Yan, M.G., Liu, B.C., LI, J.G., 2002. China Aviation Material Manual. China Standard Press, Beijing,China. Yuan, H.L., 2006. The optimal design of profiled ring blank and rolling process numerical simulation. Dorctoral Dissertation for Huazhong University of Science and Technology Wuhan,China.
291