Ken M. Wallace

A Methodology for Creating Ontologies for Engineering Design

This paper describes a methodology for developing ontologies for engineering design. The methodology combines a number of methods from social science and computer science, together with taxonomies developed in the field of engineering design. A case study is used throughout the paper focusing upon the use of an ontology for searching, indexing and retrieving of engineering knowledge. An ontology for indexing design knowledge can assist the users to formulate their queries when searching for engineering design knowledge. The root concepts of the ontology were elicited from engineering designers during an empirical research study. These formed individual taxonomies within the ontology and were validated through indexing a set of ninety-two documents. Relationships between concepts are extracted as the ontology is populated with instances. The identified root concepts were found to be complete and sufficient for the purpose of indexing. A thesaurus and an automatic classification are being developed as a result of this evaluation. The methodology employed during the test case is presented in this paper. There are six separate stages, which are presented together with the research methods employed for each stage and the evaluation of each stage. The main contribution of this research is the development of a methodology to allow researchers and industry to create ontologies for their particular purpose and to develop a thesaurus for the terms within the ontology. The methodology is based upon empirical research and hence, focuses upon understanding a user’s domain models as opposed to extracting an ontology from documentation.Copyright

DRed and Design Folders: A Way of Capturing, Storing and Passing On Knowledge Generated During Design Projects

This paper describes a software tool called DRed (the D esign R ationale ed itor), that allows engineering designers to record their design rationale (DR) at the time of its generation and deliberation. DRed is one of many proposed derivatives of the venerable IBIS concept, but by contrast with other tools of this type, practicing designers appear surprisingly willing to use it. DRed allows the issues addressed, options considered, and associated arguments for and against, to be captured graphically. The software, despite still being essentially a research prototype, is already in use on high profile design projects in an international aerospace company, including the presentation of results of design work to external customers. The paper compares DRed with other IBIS-derived software tools, to explain how it addresses problems that seem to have made them unsuitable for routine use by designers. In addition to the capture and presentation of the DR itself, the set of linked DR graphs can be used to provide a map of the contents of an electronic Design Folder, containing all the documents created by an individual or team during a design project. The structure of the knowledge model instantiated in such a Design Folder is described. By reprising a design case study published at the DTM 2003 conference, concerning the design of a Mobile Arm Support (MAS), the DRed knowledge model is compared with the previously proposed Design Data Model (DDM), to show how it addresses the shortcomings identified in the DDM. Finally the methodology and results of the preliminary evaluation of the use of DRed by aerospace designers are presented.Copyright

Identifying and supporting the knowledge needs of novice designers within the aerospace industry

A long-term research project is being undertaken to understand how to support engineering designers through the provision of appropriate knowledge during the design process. Empirical research already carried out within the aerospace industry found that novice designers were aware of their knowledge needs in only 35% of their queries. In addition, observations of experienced and novice designers identified design strategies employed by experienced designers. The aim of this research was to develop and evaluate a method of support for designers based upon these findings. This paper describes the development of the method and its preliminary evaluation within the aerospace industry. The evaluation indicated that the method was able to reduce the number of inappropriate questions asked by novice designers.

Understanding how the information requests of aerospace engineering designers influence information-seeking behaviour

Engineering design activities rely heavily on the richness of knowledge and information available. In order to progress their tasks, designers undertake numerous searches to answer their information requests. A review of the literature on information seeking (IS) found that the role of the types of information request on source selection has been little investigated compared with that of the source itself, the seeker, the seeker–source relationship and the work context. This paper describes the empirical research that was carried out in collaboration with the aerospace group of a major power systems company to understand the relationship between information searches and information requests raised by designers involved in individual design tasks. The empirical part of the research consisted of participating in a design activity undertaken within the collaborating company and in conducting analytical studies with its designers. Two large data sets were gathered using a diary study and observations with shadowing. Based on the literature review and empirical research, a framework to describe how engineering designers access information was developed. The context of an information request, the request itself and the associated search were characterised through three groups of categories. Designers were found to raise requests to acquire information and to process information. The main sources used to answer these requests were colleagues, databases and drawings. In examining how sources are used, the research confirmed that designers generally prefer to source knowledge and information through informal interactions with their colleagues and demonstrated that colleagues are more often consulted to answer the requests to process information and documentary sources to answer the requests to acquire information. The results on the use of the media to communicate with colleagues indicated that face-to-face interactions are preferred to answer the requests to process information and phone calls to answer the requests to acquire information. Overall, the research emphasises the importance of unplanned face-to-face interactions in answering complex requests. Interpersonal communication emerges as a key resource for progressing design problems in a practical way, learning how to design and establishing social networks. For researchers, this study extends the existing knowledge by proposing to analyse and interpret engineering designers’ IS behaviour based on their information needs. Practitioners from current manufacturing organisations can use these findings to evaluate existing knowledge management efforts and subsequently undertake new interventions to improve the management of their knowledge and information.

Decision-making in Engineering Design: Observations from Design Experiments

SUMMARY During the engineering design process, decisions influence both the direction of the process and the definition of the product. The process of decision-making is not understood completely, so more research into how decisions are actually made is needed. To that end, descriptive studies, i.e. studies on how design processes actually take place, have been undertaken in a controlled laboratory environment at Cambridge Engineering Design Centre. This paper describes these studies, discusses the main research issues, presents some results and identifies some of the limitations of this type of research.

Evaluating a functional basis

One of the motivations to develop a standard language to describe the functions of a product or artifact is to facilitate the indexing, search and retrieval of design knowledge. Taxonomies describing a standard language of functions have been developed by a number of researchers. This paper describes an empirical study carried out to evaluate two such taxonomies for their suitability to indexing design knowledge. The results of eighteen interviews conducted into two separate aerospace companies in the UK are presented. Engineering designers were asked to describe the function of assemblies or components that they were familiar with. In total, 207 descriptions of functions were collected. These descriptions have been analysed and compared to the two taxonomies evaluated. This paper describes the findings from these studies and draws conclusions on the suitability of these taxonomies for indexing design knowledge.Copyright

Supporting the Knowledge Life-Cycle

This paper describes the mismatch between the information sources available to designers and the information sources accessed by designers. Possible reasons for this mismatch are accessibility, availability and trustworthiness. Current support for product development does not deal with this issue because only part of the knowledge life-cycle is addressed. A model of the knowledge life-cycle is introduced that focuses on knowledge generation. To support the knowledge life-cycle,· a system is introduced based on a model of the design process. This system, PROS US, is discussed and the issues needing further development identified.

The introduction of a design heuristics extraction method

Abstract This paper introduces a novel method for analyzing conceptual design data. Given a database of previous designs, this method identifies relationships between design components within this database. Further, the method transforms these relationships into explicit design knowledge that can be used to generate a ‘heuristic-based’ model of the design domain for use at the conceptual stage. This can be viewed as a knowledge extracting method for the conceptual design stage. Such a method is particularly interesting, as the conceptual stage typically lacks explicit models to describe the trade-offs that must be made when designing. The method uses either principal components analysis or self-organizing maps to identify the relationships, and this paper describes all the elements required by the method to successfully extract and verify design knowledge from design databases.

Design-led component selection

Designers are under increasing pressure to deliver better, cheaper products in less time. To the designer of low-volume or one-off products, using past designs and catalogued components is often essential. Little software support exists early in the design process for designers wanting to use catalogued components. Most information on off-the-shelf solutions has to be gleaned from component manufacturers catalogues. Searching a large number of paper- or computer-based catalogues soon becomes unwieldy and inefficient. This paper describes a method to assist designers in selecting components early in the design process.

Characterising in Detail the Information Requests of Engineering Designers

This paper describes empirical research that was carried out in collaboration with the aerospace group of a major power systems company to characterise in detail the information requests of engineering designers. The research started by reviewing current classifications of information requests as well as the problem solving and reasoning processes. The data collection consisted of integrating ethnographical participation with a diary study and observations with shadowing. Two large data sets were gathered from the diary study and the observations. The literature review and the studies conducted in this project led to the development of six categories to classify information requests, namely objective, subject, response process, response type, direction of reasoning and behaviour type. The development of the categories was data driven. Each of the first four categories was designed to classify one characteristic of information requests. The final two categories were designed to be used together and to classify one additional characteristic of information requests. The paper discusses in detail the categories. The results of the quantitative analysis of the two data sets using the six categories are also presented. At a later stage of the research, a coding scheme to classify multiple characteristics of information requests was extracted from the six categories. The paper discusses the scheme with respect to current classifications. The implications of the research results to improve the support that can be provided to designers are also discussed and presented.Copyright