Carnegie Mellon University
Browse
file.pdf (1.27 MB)

A representation for algorithmic expansion of engineering designs

Download (1.27 MB)
journal contribution
posted on 1992-01-01, 00:00 authored by Vital Aelion, Jonathan Cagan, Gary J. Powers, Carnegie Mellon University.Engineering Design Research Center.
Abstract: "A knowledge representation based on first principles and a library of techniques for expanding the space of design alternatives are reviewed. They have been developed for the algorithmic innovation of engineering designs. The knowledge representation consists of an algebraic design objective, a set of algebraic equality and inequality constraints, and variable and constraint typing specifications. The algebraic relations model the fundamental physical and chemical phenomena, as well as manufacturing, safety, and performance considerations. Within this representation, variable typing describes physical, processing and manufacturing characteristics of variables within a design. Constraint typing describes the region(s) in which each constraint is applicable in a design. This representation may be useful in applications which are amenable to mathematical representations, and require additional information on variable characteristics and constraint applicability, to achieve their reasoning goals. The expansion techniques introduce new independently modeled features in a known design. Presently there are two expansion techniques: Dimensional Variable Expansion, DVE, and Input Variable Expansion, IVE. These techniques are the means for modifying design topology. DVE and IVE have been defined as mathematical operators, to manipulate our knowledge representation. The concept of design space expansion is useful independently of knowledge representation. The knowledge representation and the library of design space expansion techniques have been combined into the 1stPRINCE design methodology for innovation in engineering design. This methodology uses optimization information to decide which expansion technique may produce improved designs and induction to examine limiting behavior. Starting from an initial design, 1stPRINCE has algorithmically innovated interesting engineering concepts, such as the electric bus, the tapered and hollow beam, the wheel, and the plug flow reactor."

History

Publisher Statement

All Rights Reserved

Date

1992-01-01

Usage metrics

    Exports

    RefWorks
    BibTeX
    Ref. manager
    Endnote
    DataCite
    NLM
    DC