Access Type

Open Access Dissertation

Date of Award

January 2012

Degree Type

Dissertation

Degree Name

Ph.D.

Department

Industrial and Manufacturing Engineering

First Advisor

Kyoung-Yun J. Kim

Abstract

Modern product innovation is often delivered by fusing technologies from different domains. To meet the current speedy pressure of product innovation, the conceptual design is a very critical stage to develop an innovative product. This research presents a new design paradigm, Transformative Product Design (TPD), to meet the demands in the conceptual design. TPD aims to design a new product from a combination out of a base product and reference products, which have been developed.

To expedite the TPD paradigm, Interaction Network is introduced to represent a product design by following a representative product design representation method in the teleological perspective. However, generating an interaction network of a product is very cumbersome and somewhat subjective. In the past decade, a few research works have been reported to identify functions of a product in a systematic way. They are not satisfactory to develop an interaction network for TPD systematically. To systematically generate an interaction network, this research adopts functional requirements in design documents as the source of the proposed method. The proposed method in this research aims to identify functions, structures and dynamic behaviors between structures from the functional descriptions in natural (English) language by adapting natural language-based object identification methods in the software engineering.

Another aim of this research is to provide semantic capability to utilize the use of a natural language throughout the design transformation process, taking into account cross-disciplinary design knowledge that domain experts could struggle if the knowledge is beyond their expertise. To fulfill the demand, this research builds research methods for the design transformation, based on the proposed similarity score, i.e., Stem-POS based Similarity Score. By employing the proposed score, this research detects concept functions and generates transformative design alternatives with interaction networks. Additionally, for the design transformation, the degree of transformability is to check how much products are semantically related each other, by adapting the semantic similarity score.

Finally the proposed methods have been implemented in a Computer-Aided Transformative Design (CATPD) supporting system with minor human involvement to a minimum. The proposed methods and the system are validated according to Stieglitz's convergence types.

Download

Share

COinS