4.6 Article

A new adaptive multiscale method based on the estimate of residual forces for static analysis of heterogeneous materials

Journal

FINITE ELEMENTS IN ANALYSIS AND DESIGN
Volume 101, Issue -, Pages 34-45

Publisher

ELSEVIER SCIENCE BV
DOI: 10.1016/j.finel.2015.04.001

Keywords

Heterogeneous materials; Adaptive multiscale method; Multi-node extended multiscale finite; element method (multi-node EMsFEM); Residual force; Multiscale computation

Funding

  1. National Natural Science Foundation of China [11402178, 11472196]
  2. China Postdoctoral Science Foundation [2014M552078, 2014M562055]
  3. Hubei Provincial Natural Science Foundation [2014CFB336]
  4. Open Research Fund Program of Wuhan University [2042014kf0002]

Ask authors/readers for more resources

A new adaptive multiscale method (AMM) is developed based on the estimate of residual forces for static analysis of heterogeneous materials. The AMM is established by combining multi-node extended multiscale finite element method (multi-node EMsFEM) with a new proposed macroscopic node adaptive algorithm. In our previous multiscale computations, macroscopic nodes are placed uniformly along each edge of multi-node coarse element without considering local strain or displacement gradient. In this paper, to optimize the distribution of macroscopic nodes, a new adaptive algorithm is proposed based on the estimate of residual forces. Numerical experiments have indicated that residual forces exist even for linear elastic problems. For boundary external loading cases, residual forces only exist on the edges of coarse element. Besides, computations indicate that residual forces can reflect local relative errors in the multi-node EMsFEM computations. Thus it is reasonable and suitable to take residual forces as local relative error indicators in the multi-node EMsFEM computations. Finally, the AMM is developed based on this idea. To verify the validity of this proposed method, three typical numerical examples are carried out The examples demonstrate that nearly optimal distributions of macroscopic nodes can be obtained by employing the proposed AMM. (C) 2015 Elsevier B.V. All rights reserved.

Authors

I am an author on this paper
Click your name to claim this paper and add it to your profile.

Reviews

Primary Rating

4.6
Not enough ratings

Secondary Ratings

Novelty
-
Significance
-
Scientific rigor
-
Rate this paper

Recommended

No Data Available
No Data Available