An object-oriented application framework for finite element analysis in structural engineering
Abstract
This thesis investigates a means for applying the object-oriented paradigm to the domain of finite element analysis. The thrust is to investigate, design, implement, and demonstrate FE++, an extensible object-oriented application framework to support the construction of a wide variety of FEA application programs. The major goal of FE++ is to support research in a broad domain of finite element analysis by providing an object-oriented architecture and a set of programming abstractions that offer "test-bed" capability for rapid prototyping of a variety of FEA applications. These target applications range from simple linear elastic analysis to fully nonlinear transient analysis. The abstractions and mechanisms provided by FE++ ease the configuration of different solution algorithms. The FE++ architecture accommodates finite element models consisting of different types of elements and variable numbers of degrees of freedom per node. No limit is imposed on the number of degrees of freedom per node. Furthermore, FE++ supports dynamically evolving topology, therefore facilitating construction of FEA application programs to analyze structures whose topology changes during analysis. The FE++ architecture seamlessly integrates under a unified framework the support for linear and nonlinear analysis, and static and transient analysis. The thesis describes the design and implementation of the FE++ architecture. The application of FE++ is demonstrated with several prototype FEA applications, including: (1) analysis of a structure consisting of heterogeneous elements (2) nonlinear analysis involving large displacements, and (3) analysis of fragmentation simulation under high-rate transient loads. From these examples and from the presentation of the FE++ architecture and component classes, it can be inferred that FE++ can reduce substantially the implementation time and effort when compared with conventional tools, and that FE improves the clarity, expressiveness, and reliability of client code.
Degree
Ph.D.
Advisors
Chen, Purdue University.
Subject Area
Civil engineering|Mechanical engineering|Computer science
Off-Campus Purdue Users:
To access this dissertation, please log in to our
proxy server.