Digital transformation supported by Knowledge-Based Engineering and semantic modeling for automating engineering tasks

Abstract

The culture of product design is shifting from case-by-case development to the Knowledge-Based (KB) paradigm. This shift aims to foster knowledge sharing and reuse across different stages and groups in engineering. Within this context, engineering knowledge encompasses both product data and design rules. The paradigm shift requires the digital transformation of product design from document-centric to knowledge-centric paradigm. Despite existing research primarily focusing on product data representation, there remains a notable gap in addressing a comprehensive framework formalizing design knowledge, particularly in terms of representing various types of design rules. Knowledge-Based Engineering (KBE) is a technology that employs knowledge representation languages for describing facts and rules, with the goal of automating engineering tasks through reasoning abilities, including design, analysis, and optimization. KBE often integrates with a CAD kernel for geometry manipulation, and its explicit product data representation facilitates data exchange with Computer-Aided Analysis (CAA) tools, enhancing its data processing and computation capabilities. The semantic web is designed to provide a common framework for sharing and reusing data across various sources. It aims to transform the existing “web of documents” to be a “web of data”, aligning with the objectives of KBE paradigm. Consequently, the semantic web stack can be utilized to represent product data and universal design rules. Leveraging the semantic representation of design knowledge, various applications can be developed to encapsulate case-specific rules. In summary, KBE and semantic modeling offer a potential framework for the digital transformation of the product design paradigm. This thesis aims to propose a practical framework that leverages KBE and se mantic modeling. This framework is designed to facilitate the transition from a document-centric paradigm to a knowledge-based paradigm in engineering design. A key feature of this framework is its ability to capture and formalize designers’ intent, enabling the rapid generation of product variants in response to changes in intent. Compared with hard-coded ad-hoc software applications, this framework offers better interoperability and extendability. Consequently, the integration of various digital tools for diverse engineering tasks is facilitated, supporting longterm knowledge reuse. The exploration of interaction with Natural Language Processing (NLP) tools for user-friendly workflow composition highlights its promising potential. The thesis is presented as a collection of publications with the goal of facilitating the transition towards a knowledge-based paradigm in engineering design. The structure of the thesis consists of five chapters: Chapter 1 provides a comprehensive overview of the background and the research questions that underpin this Ph.D. work. Chapter 2 delves into the theoretical framework and examines related works that have preceded this research. Chapter 3 summarizes the primary contributions of each individual research paper. Chapter 4 elaborates the connections between the individual research papers and the three derived research questions. Finally, Chapter 5 concludes the thesis and provides perspectives for future research.