Advances in Engineering Software (1978)最新文献

A hierarchical boundary model of three dimensional objects provides a representation of the surfaces enclosing it at successively finer levels of specification. A hierarchical graph structure for encoding the boundary of a solid object at variable resolution is described. This model, called structured edge-face graph, is based on a face oriented relational description of the object boundary. In the paper, the concept of variable resolution surface representation is briefly reviewed, and the structured edge-face is described in detail. A data structure for encoding a structured edge-faced graph is presented. The concept of adjacency relation between pairs of primitive boundary entities at fixed accuracy is introduced, and adjacency finding algorithms, which operate on a structured edge-face graph object representation, are presented. Four basic operations for defining and manipulating a structured model are described. These primitives support both a bottom-up and a top-down development of the hierarchical boundary model.

A simple FORTRAN routine, employing a combination of third-order Schroder iteration and Newton-Raphson method, is given for computing gamma percentiles

An implementation of the well-known LU decomposition method for solving systems of linear equations is presented. Operation counts typically stated for this method are derived from a theoretical analysis of the algorithm and ignore the practical aspects of the implementation. The overhead associated with assessing elements of the two-dimensional coefficient array are explored herein, and seen to be substantial. This overhead is incorporated into the analysis of LU decomposition and new operation counts are developed. By exploiting the addressing capabilities of C the cost of array access is significantly reduced, resulting in an efficient implementation. The techniques employed here can be applied to a wide variety of C programs which utilize multi-dimensional arrays.

The simulation of reinforced concrete frames by networks, with bars obeying uniaxial stres-strain laws of concrete or steel, is proposed. Formulae for the determination of concrete bar sections are derived. Concrete σ-ε law, including cracking and plastic behavior, is described by 3 constitutive variables; steel σ-ε law, including plastic behavior, is described by 1 constitutive variable. A simple program is presented for the nonlinear static analysis of such network models based on the incremental loading technique. This program is used for the analysis of a plane, one story reinforced concrete frame under cyclic horizontal loading of its girder, for which experimental data are available. The computational results are found in good agreement with the experimental ones.

The INTERP free format data interpretation subroutine system for lexical analysis is described in this paper. This system, written in ANSI standard FORTRAN 77, is flexible, easy to use, small, and efficient. These qualities make the addition of INTERP to an application program quite simple. INTERP has been used to interpret problem oriented languages for several finite element programs operating both in batch and interactive environments on micro, supermini, and supercomputers. While INTERP is presented here in the context of translating problem oriented languages, the system has been used in several other applications including: a low-level assembler, a data manipulator, and a modular, interactive user interface system with integrated parsing and prompting.

The application of rule-based programming, commonly associated with expert systems, to programming tasks normally associated with procedural software is considered. To illustrate the ideas, a toy finite element package has been developed, consisting of a set of processors which are configured at run time by interpreting a set of rules. A typical rule relates the immediate prerequisites for a particular processor to its output. The algorithm used to interpret the rules and produce a linearised graph listing the order of execution of processors, is described. The potential of the methodology is to achieve a package that can reconfigure itself to meet a user's request. It would avoid redundant computation and be easy to modify. Issues in data management that arise from adoption of the methodology are briefly discussed, including the design of a suitable database.