资源类型

期刊论文 4

年份

2022 1

2008 1

2006 1

2000 1

关键词

函数逼近能力 1

计算能力 1

过程神经元网络 1

连续性 1

检索范围:

排序: 展示方式:

Implicit Heaviside filter with high continuity based on suitably graded THB splines

《机械工程前沿(英文)》 2022年 第17卷 第1期   页码 14-14 doi: 10.1007/s11465-021-0670-2

摘要: The variable density topology optimization (TO) method has been applied to various engineering fields because it can effectively and efficiently generate the conceptual design for engineering structures. However, it suffers from the problem of low continuity resulting from the discreteness of both design variables and explicit Heaviside filter. In this paper, an implicit Heaviside filter with high continuity is introduced to generate black and white designs for TO where the design space is parameterized by suitably graded truncated hierarchical B-splines (THB). In this approach, the fixed analysis mesh of isogeometric analysis is decoupled from the design mesh, whose adaptivity is implemented by truncated hierarchical B-spline subjected to an admissible requirement. Through the intrinsic local support and high continuity of THB basis, an implicit adaptively adjusted Heaviside filter is obtained to remove the checkboard patterns and generate black and white designs. Threefold advantages are attained in the proposed filter: a) The connection between analysis mesh and adaptive design mesh is easily established compared with the traditional adaptive TO method using nodal density; b) the efficiency in updating design variables is remarkably improved than the traditional implicit sensitivity filter based on B-splines under successive global refinement; and c) the generated black and white designs are preliminarily compatible with current commercial computer aided design system. Several numerical examples are used to verify the effectiveness of the proposed implicit Heaviside filter in compliance and compliant mechanism as well as heat conduction TO problems.

关键词: topology optimization     truncated hierarchical B-spline     isogeometric analysis     black and white designs     Heaviside filter    

Meshless numerical method based on tensor product

SUN Haitao, WANG Yuanhan, MIAO Yu

《结构与土木工程前沿(英文)》 2008年 第2卷 第2期   页码 166-171 doi: 10.1007/s11709-008-0021-y

摘要: A normalized space constructed by tensor product is used in field function approach to give a special case of moving least squares (MLS) interpolation scheme. In the regular domain, the field function which meets homogenous boundary conditions is constructed by spanning base space to make the MLS interpolation scheme simpler and more efficient. Owing to expanded basis functions selection, some drawbacks in general MLS method, for example repeated inversion, low calculation efficiency, and complex criterions, can be avoided completely. Numerical examples illustrate that the proposed method is characterized by simple mathematical concept, convenient repeat calculations with high accuracy, good continuity, less computation and rapid convergence.

关键词: mathematical     continuity     inversion     selection     efficient    

Principle and curve of planetary indexing cam mechanisms

LIU Ming-tao, ZHANG Ce, YANG Yu-hu

《机械工程前沿(英文)》 2006年 第1卷 第4期   页码 475-478 doi: 10.1007/s11465-006-0063-6

摘要: A new type of planetary indexing cam mechanism called type !c is presented. The transmission principle of type !c is discussed. Based on the principle of formation of conjugate tooth contour, the equations of the cam s pitch curve are derived. The continuity and symmetry of the cam s pitch curve are also analyzed. It can be concluded that: 1) This type has no output institution so the output precision can be improved; 2) The asymmetrical two-ring structure is used to overcome dead point and its phase angle is equal to (π - 2π/n); 3) The continuous condition of the cam s pitch curve is that the average transmission ratio is -n in an intermittent motion period.

关键词: continuity     planetary     average transmission     symmetry     intermittent    

过程神经元网络的若干理论问题

何新贵,梁久祯

《中国工程科学》 2000年 第2卷 第12期   页码 40-44

摘要:

文章提出一种过程神经元模型,其输人为与时间有关的函数或过程,它是传统人工神经元模型在时间域上的扩展。基于这种过程神经元模型,给出了一种仅含一个隐层的前馈型过程神经网络模型,即基展开过程神经元网络模型。证明了相应的连续性定理,逼近定理,计算能力定理等。

关键词: 过程神经元网络     函数逼近能力     计算能力     连续性    

标题 作者 时间 类型 操作

Implicit Heaviside filter with high continuity based on suitably graded THB splines

期刊论文

Meshless numerical method based on tensor product

SUN Haitao, WANG Yuanhan, MIAO Yu

期刊论文

Principle and curve of planetary indexing cam mechanisms

LIU Ming-tao, ZHANG Ce, YANG Yu-hu

期刊论文

过程神经元网络的若干理论问题

何新贵,梁久祯

期刊论文