|
| Factors affecting computational precision of MLS-based meshless method |
| Revised:July 06, 2001 |
| View Full Text View/Add Comment Download reader |
| DOI：10.7511/jslx20033061 |
| KeyWord:meshless method,weight function,compact support,computational precision,MLS |
| Lou Luliang \ Zeng Pan |
|
| Hits: 3851 |
| Download times: 8 |
| Abstract: |
| The computational precision of MLS\|based meshless method is affected by many factors besides the nodal distribution density and the order of the basis, some of which are the weight functions, the compact support's size, and how to enforce the essential boundary conditions. In this paper three commonly used weight functions, exponential function, spline function, and circular function, were analyzed including the approximation precision, the convergence behavior, and the computational efficiency. At the same time the effect of the compact support and the enforcement of essential boundary conditions on the computational precision was studied. Through analysis how to choose the weight function and the compact support was given. When the constrained DOF was relatively more, the enforcement of the essential boundary conditions by collocation method would cause the results oscillation. A stabilization term was introduced to improve the precision on the boundary, which was validated by the example of a rectangular plate with centered hole under uniform tensile loading. This paper concludes with a sample stress analysis of J23\|10 press's body, and the numerical results are in good agreement with the experimental results. |
|
|
|