日本語フィールド
著者：Hagihara, Seiya; Motoda, Shinji; Tsunori, Mitsuyoshi; Ikeda, Toru; Miyazaki, Noriyuki題名：Adaptive meshfree method based on adding and moving nodes発表情報：International Journal of Computational Methods 巻: 3 号: 4 ページ: 429 - 443キーワード：概要：The objective of the present research is to apply element-free Galerkin (EFG) method to adaptive analyses. It is necessary to estimate error of numerical solutions of EFG method for adaptive analyses to evaluate accuracy of EFG method. Posteriori errors can be estimated by differences of solutions between the linear and the quadratic basis functions. But it is not economical to perform the two respective calculations in regard to the linear and the quadratic basis function. Then the linear function is used only when the stiffness matrix is calculated. Then both the linear and the quadratic basis function are used when the stress and strain are calculated. The error estimation is performed in an each background cell by using the error of energy norm. In the adaptive analysis, a node is added at the quadrature point which is the same as the center of gravity of a background cell where the error is higher than the threshold value. The nodal relocation method is applied to smoothing the distribution of nodes in domain of an analysis model. The nodal relocation method in which nodes are automatically moved and relocated using physical interbubble forces called bubble meshing for FEM is applied to the adaptive analysis after additional nodes are generated. The nodes are relocated corresponding to the error of the background cells. The calculations of the analysis can be repeated again after the posteriori error estimation. The adaptive EFG method and the nodal relocation method are applied to a problem of an infinite plate with a hole subjected to a uniform tension. Nodal density is increased at the vicinity of the hole where the error is large in the analysis. The nodal relocation method can be successful to relocate the nodes which are generated at the quadrature points of higher posteriori error. The difference between the calculated solution and the exact solution are smaller than that of the previous solution as the calculations are repeated. © World Scientific Publishing Company.抄録：英語フィールド
Author：Hagihara, Seiya; Motoda, Shinji; Tsunori, Mitsuyoshi; Ikeda, Toru; Miyazaki, NoriyukiTitle：Adaptive meshfree method based on adding and moving nodesAnnouncement information:International Journal of Computational Methods Vol: 3 Issue: 4 Page: 429 - 443An abstract：The objective of the present research is to apply element-free Galerkin (EFG) method to adaptive analyses. It is necessary to estimate error of numerical solutions of EFG method for adaptive analyses to evaluate accuracy of EFG method. Posteriori errors can be estimated by differences of solutions between the linear and the quadratic basis functions. But it is not economical to perform the two respective calculations in regard to the linear and the quadratic basis function. Then the linear function is used only when the stiffness matrix is calculated. Then both the linear and the quadratic basis function are used when the stress and strain are calculated. The error estimation is performed in an each background cell by using the error of energy norm. In the adaptive analysis, a node is added at the quadrature point which is the same as the center of gravity of a background cell where the error is higher than the threshold value. The nodal relocation method is applied to smoothing the distribution of nodes in domain of an analysis model. The nodal relocation method in which nodes are automatically moved and relocated using physical interbubble forces called bubble meshing for FEM is applied to the adaptive analysis after additional nodes are generated. The nodes are relocated corresponding to the error of the background cells. The calculations of the analysis can be repeated again after the posteriori error estimation. The adaptive EFG method and the nodal relocation method are applied to a problem of an infinite plate with a hole subjected to a uniform tension. Nodal density is increased at the vicinity of the hole where the error is large in the analysis. The nodal relocation method can be successful to relocate the nodes which are generated at the quadrature points of higher posteriori error. The difference between the calculated solution and the exact solution are smaller than that of the previous solution as the calculations are repeated. © World Scientific Publishing Company.