Intelligent materials selection inconceptural design

Dissertant

Abbas, Suad Hamzah

University

University of Technology

Faculty

-

Department

Department of Production Engineering and Metallurgy

University Country

Iraq

Degree

Ph.D.

Degree Date

2006

English Abstract

Materials selection is an integral part of the design of a product.

The needs to combine selection of materials (SM) processes during the early stages of the design have previously been realized.

The work documented in this dissertation is an attempt to ensure that there is no disconnect between function oriented design and the material that are applicable to that design.

Since the information in the conceptual stage is fuzzy and ambiguous and properties are dealt within ranges, which make the selection process very difficult.

The work proposes methodology which helps a designer to select the best sub-group material that satisfy the design requirements in the early design stages which help to move the decisions on selection of materials to the explore conceptual solution.

The developed methodology serves the concurrent engineering (CE) Philosophy to take the correct decision early in the design process.

This leads in avoiding additional cost, the transient mutations in manufacturing stage and improving the quality of product.

The developed methodology consists of two concurrent SM methods ; the first one is neural network selection of materials method which helps the designer in selecting correct sub-group materials when the design concept is in its beginning.

Neural network tool was being used to resemble the human's intuitive ability in solving problems though learning.

An algorithm was being built to perform SM where design is still a concept and only described verbally.

The second one was concurrent qualitative selection of materials method (CQM).

The method is modified from quantitative method called weighted property method used for selecting materials in the detailed stage.

The method was modified to qualitative method, which means that the input data for the materials property of the design must be qualitative data which consisting of groups or sub-groups material, range value properties-and approximate values.

Two case studies (turbine blades and aircraft skin) were examined and satisfactory results were obtained.

The results of the second developed method supports the results obtained from the first method but in a more accurate manner.

This is because the neural network environment resembles the human approach in selecting material in the conceptual stage.

Main Subjects

Engineering & Technology Sciences (Multidisciplinary)

Topics

• Production engineering
• Smart materials

• APA
• MLA
• AMA

Language

English

Data Type

Arab Theses

Record ID

BIM-305819