A novel method for prediction of truss geometry from topology optimization

详细信息    查看全文

• 作者：A. Mandhyan ; Gaurav Srivastava ; S. Krishnamoorthi
• 关键词：Topology optimization ; Truss ; like structures ; Grayscale image ; Optimal truss shape
• 刊名：Engineering with Computers
• 出版年：2017
• 出版时间：January 2017
• 年：2017
• 卷：33
• 期：1
• 页码：95-106
• 全文大小：
• 刊物类别：Computer Science
• 刊物主题：Computer-Aided Engineering (CAD, CAE) and Design; Math. Applications in Chemistry; Systems Theory, Control; Calculus of Variations and Optimal Control; Optimization; Classical Mechanics; Appl.Mathemat
• 出版者：Springer London
• ISSN：1435-5663
• 卷排序：33

文摘

An important task in designing a truss structure is to determine the initial configuration of the truss. In the absence of an efficient optimization technique, the selection of initial geometry is based on a trial and error procedure or standard truss configurations or past experiences. In this work, a fully automated algorithm is proposed which can be used to predict initial truss geometry from the grayscale images obtained from topology optimization of design domain. It predicts the locations of joints and the connectivity of members. It also estimates the approximate cross-sectional areas of the members. The interpreted truss geometry can be modified or directly used for structural analysis and design. Numerical examples are presented to demonstrate the functioning of the algorithm under various scenarios. The developed algorithm has been implemented as part of a web-based truss design application, previously developed by the same authors.