An automated approach to quantifying functional interactions by mining large-scale product specification data. Issue 1 (3rd March 2016)

• Record Type: Journal Article
• Authors: Kang, Sung Woo; Tucker, Conrad
• Is Part Of: Journal of engineering design. Volume 27:Issue 1/3(2016)
• Page Start: 1
• Page End: 24
• Publication Date: 2016-03-03
• Subjects: Functional interaction -- text mining -- design structure matrix; Engineering design -- Periodicals; Design, Industrial -- Periodicals; Industrial engineering -- Periodicals; 620.0042
• DOI: 10.1080/09544828.2015.1083539
• Languages: English
• ISSNs: 0954-4828

Abstract

ABSTRACT: The authors of this work hypothesise that the semantic relationship between modules' functional descriptions is correlated with the functional interaction between the modules. A deeper comprehension of the functional interactions between modules enables designers to integrate complex systems during the early stages of the product design process. Existing approaches that measure functional interactions between modules rely on the manual provision of designers' expert analyses, which may be time consuming and costly. The increased quantity and complexity of products in the twenty-first century further exacerbates these challenges. This work proposes an approach to automatically quantify the functional interactions between modules, based on their textual technical descriptions. Compared with manual analyses by design experts who use traditional design structure matrix approaches, the text-mining-driven methodology discovers similar functional interactions, while maintaining comparable accuracies. The case study presented in this work analyses an automotive climate control system and compares the functional interaction solutions achieved by a traditional design team with those achieved following the methodology outlined in this paper.