A multi-level refinement adaptive scheme with high efficiency and accuracy
Abstract
Purpose
Traditional adaptive analysis based on a coarse mesh, using finite element method (FEM) analysis, produces the original solution. Then post-processing the result and figuring out the regions should be refined and these regions refined once. Finally, this new mesh is used to get the solution of first refinement. After several iterations of above procedures, we can achieve the last result that is closer to the true solution, which takes time, making adaptive scheme inpractical to engineering application. The paper aims to discuss these issues.
Design/methodology/approach
This paper based on FEM proposes a multi-level refinement strategy with a refinement strategy and an indicator. The proposed indicator uses value of the maximum difference of strain energy density among the elements that associated with one node, and divides all nodes into several categories based on the value. A multi-level refinement strategy is proposed according to which category the node belongs to refine different elements to different times rather than whether refine or not.
Findings
Multi-level refinement strategy takes full use of the numerical calculation, resulting in the whole adaptive analysis that only need to iterate twice while other schemes must iterate more times. Using much less times of numerical calculation and approaches, more accurate solution, making adaptive analysis more practical to engineering.
Originality/value
Multi-level refinement strategy takes full use of the numerical calculation, resulting in the whole adaptive analysis only need iterate twice while other schemes must iterate more times. using much less times of numerical calculation and approaches more accurate solution, making adaptive analysis more practical to engineering.
Keywords
Citation
Ding, C., Cui, X., Li, C., Li, G. and Wang, G. (2016), "A multi-level refinement adaptive scheme with high efficiency and accuracy", Engineering Computations, Vol. 33 No. 7, pp. 2216-2236. https://doi.org/10.1108/EC-05-2015-0119
Publisher
:Emerald Group Publishing Limited
Copyright © 2016, Emerald Group Publishing Limited