TY - JOUR
T1 - Acceleration strategies based on bubble-type adaptive mesh refinement method
AU - Guo, Wei
AU - Nie, Yufeng
AU - Zhang, Weiwei
N1 - Publisher Copyright:
© 2019 International Association for Mathematics and Computers in Simulation (IMACS)
PY - 2020/4
Y1 - 2020/4
N2 - Although the adaptive mesh refinement method based on bubble placement can generate a high-quality mesh, the efficiency of bubble placement method remains to be improved. In this study several acceleration strategies are proposed to reduce the cost of simulation. The specific strategies are given as follows: in order to reduce the number of simulation rounds, the bubbles at the new refinement level are added in advance according to the estimated bubble distribution; multilevel time step is adopted for time integration and the time step is defined separately for moving bubbles and oscillating bubbles; a scheme for setting cell length adaptively is proposed and the cell searching method is modified to improve the efficiency of establishing adjacent list; efficiency of updating bubble size is also enhanced by promoting the localization process. Numerical examples illustrate that the computing cost can significantly decrease by 70% via adopting the strategies above, while keeping the mesh quality unchanged compared with the traditional method. It shows that those strategies are efficient and suitable for adaptive mesh refinement.
AB - Although the adaptive mesh refinement method based on bubble placement can generate a high-quality mesh, the efficiency of bubble placement method remains to be improved. In this study several acceleration strategies are proposed to reduce the cost of simulation. The specific strategies are given as follows: in order to reduce the number of simulation rounds, the bubbles at the new refinement level are added in advance according to the estimated bubble distribution; multilevel time step is adopted for time integration and the time step is defined separately for moving bubbles and oscillating bubbles; a scheme for setting cell length adaptively is proposed and the cell searching method is modified to improve the efficiency of establishing adjacent list; efficiency of updating bubble size is also enhanced by promoting the localization process. Numerical examples illustrate that the computing cost can significantly decrease by 70% via adopting the strategies above, while keeping the mesh quality unchanged compared with the traditional method. It shows that those strategies are efficient and suitable for adaptive mesh refinement.
KW - Adaptive mesh refinement
KW - Algorithm efficiency
KW - Bubble packing method
KW - Multilevel adaptive algorithm
KW - Node distribution estimation
UR - http://www.scopus.com/inward/record.url?scp=85074535012&partnerID=8YFLogxK
U2 - 10.1016/j.matcom.2019.10.014
DO - 10.1016/j.matcom.2019.10.014
M3 - 文章
AN - SCOPUS:85074535012
SN - 0378-4754
VL - 170
SP - 143
EP - 163
JO - Mathematics and Computers in Simulation
JF - Mathematics and Computers in Simulation
ER -