摘要
阻碍形状优化软件广泛应用的障碍之一是依据自然设计度量描述和建立设计模型、分析模型、优化模型以及实现三个模型之间的转换.本文称这一困难为结构形状优优设计软件的适用性.本文提出了一种方法用来动态地确定平面连续体结构形状优化过程中的边界,应用基于人工智能方法的启发式规则与技术,自动生成由设计单元法表示的几何模型,也就是将一个结构自动剖分成若干个大的四边形映射单元.这些大单元对于进一步的网络生成是必要的,同时也是向全自动计算机辅助形状优化系统前进的重要一步.
One of the difficulties preventing the structural shape optimization software from wide use in industry is to describe and construct design model, optimization model and analysis model in terms of natural design variables and implemental transformations between the three models. This paper addresses the difficulty in the context of applicability of computer software. A way is presented for defining structure's boundary of plane continuum that is modified dynamically during the process of shape optimization. Heuristic rules and techniques from artificial intelligence are applied to generate the geometric model described by the design element method, i. e. , to subdivide the structure into big quadrilateral mapping elements which are necessary for further mesh generation. It is one of the important steps towards a fully automatic computer aided shape optimization system.
出处
《应用力学学报》
CAS
CSCD
北大核心
1992年第4期1-8,135,共8页
Chinese Journal of Applied Mechanics
基金
国家自然科学基金
关键词
结构形状优化
映射法
人工智能
structitral shape optimization
automatic mesh generation
mapping method
applicability of software
artificial intelligence