Strategies for the Curation of CAD Engineering Models

1. Strategies for the Curation of CAD Engineering Models 4th International Digital Curation Conference"Radical Sharing: Transforming Science?“ 1st-3rd December 2008 Edinburgh, UK Manjula Patel, Alexander Ball, Lian Ding UKOLN, DCC, IdMRC University of Bath, UK

2. Overview • Need for long-term data retention and curation in engineering • Product-Service Paradigm • Product Lifecycle Management (PLM) • Switch to creating engineering data digitally • Importance of Computer Aided Design (CAD) models • Specific curation challenges • Prototypes • LiMMA: Lightweight Models with Multilayered Annotation • RRoRIfE: Registry/Repository of Representation Information of Engineering

3. Product-Service Paradigm • Products are being purchased and sold as services rather than artefacts • Commitment to provide support for lifetime of product –could be 50 years or even longer!

4. Product Lifecycle Management • Management of product data across an (extended) enterprise over the entire lifecycle of a product • Gained prominence over the last decade or so • Globalisation has led to collaborative practices • Product development, manufacture and maintenance occur in a geographically distributed and networked environment • Product data dispersed over many organisations and locations

5. Regulators Partners Customers Design team Pre-existing information & experience Product 1 In service Upgrade In service Design team Production Product 2 Design In service Upgrade In service Production PLM Information Flows Conceptualisation; Design Organisation; Design; Evaluation; Manufacture & Delivery; Sales & Distribution; Service & Support; Decommission & Disposal

6. Product Lifecycle Management (1) • Efficient capture, representation, organisation, retrieval and reuse of product data over its entire life • Sharing of product information between many distributed organisations • Issues relating to security of information and Intellectual Property • Large variety of data, information and knowledge • Structured e.g. Geometric models and databases; test data (photographs, images, load diagrams, finite element analyses); manufacturing process plans; assembly plans; production and manufacturing logs; maintenance and service data; • Unstructured e.g. email; minutes of meetings; design reviews; contracts • Tacit knowledge e.g. design rationale; lessons learnt; procedures and processes • Exchange of information between heterogeneous computer systems • CAD/CAE/CAM(Computer Aided Design/Engineering/Manufacture) • ERP/CRM/SCM (Enterprise Resource Planning/ Customer Relationship Management/ Supply Chain Management)

7. Product Lifecycle Management (2) • Different stages of PLM require different subsets and interpretations of the same data i.e. different viewpoints e.g. Manufacturing engineers need machining features and tolerances Marketing staff need a final visualisation of the product • Data needs to be trustworthy, reliable and accurate for use in downstream processes • Vital to actively manage all information and data related to a product • Legal issues e.g. accident investigation; failure analysis; mergers and acquisitions; patent infringement • Operational support e.g. maintenance; replenishment of spare parts; recycling; disposal • Product development and management e.g. tracing design rationale; design reuse; customisation and upgrade; reverse engineering; test and validation • Various PLM systems under development (e.g. Arena, EDS, Dassault, MatrixOne, PTC etc.) • At present, no one single system capable of managing the entire product lifecycle

8. Curating CAD Model Data (1) • Computer Aided Design (CAD) models • Important in engineering practice • Definitive record of product data • Gradually superseding paper-based workflow • Communication of design to manufacturers, builders, maintenance crews and regulatory authorities • CAD software and file formats • An array of commercial CAD systems e.g. AutoCAD; CATIA; ProENGINEER; SolidWorks • Proliferation of CAD formats; proprietary, closed and subject to frequent change (unsuitable for data exchange as well as long-term preservation) • Complex dependencies and relationships between file formats, hardware and software (processing is often tied to hardware features to improve performance) • Interoperability between CAD software is virtually non-existent • Economically unviable to install and run multiple CAD systems • Product is likely to outlive several generations of the software used to design and create it!

9. CAD File Formats

10. Curating CAD Model Data (2) • CAD models comprise large volumes of data • A single, simple component can be several Megabytes • Complex products (e.g. cars, aircraft) can amount to several Gigabytes or even Terabytes • Reducing 3D model to 2D hard-copy documentation • Entails considerable information loss; also constrains reuse of data • Emulating old software and hardware • Heavy investment in IT support • Difficulties with integration into complex and more modern workflows and systems • Migration to successive proprietary formats • Risk of data loss and subtle design corruption • Rechecking and revalidating costs are substantial • Use of open standards e.g. IGES, STEP • Long timescales make it difficult to keep up with latest capabilities of CAD tools • Level of support for standards is variable between tools • [Extension of STEP for various stages of PLM underway] • Complex and dynamic environment • Difficult to reliably retrieve and trace provenance information to check reliability and quality of data • Mergers, acquisitions etc.

11. CAD Model Representations • Used in Design, Manufacture and Evaluation phases of PLM • Product information stored is still largely based on conventional product representations (geometry, topology) • Procedural definitions e.g. CSG • Explicit representation based on underlying mathematical forms e.g. B-rep • Free form surface modelling e.g. NURBS or Bézier • Feature-based (pre-defined parts) or parametric modelling • Typical components of a CAD Model • Geometry • Dimensions • Tolerances • Materials • Feature semantics • Model history

12. Challenges for CAD Models For CAD models to support information flows in PLM, they need to address: • Protection of commercially sensitive information (Intellectual Property) • Generation of viewpoint specific representations to support differing processes and users • Rapid sharing of information between geographically distributed applications and users (interoperability, platform and application independence, use of standards, reduced file sizes etc.) • Support for recording feedback from downstream processes • Curation (recording of metadata, design rationale, open formats, Representation Information, use of standards etc.)

13. Proposed Strategy To extend a CAD model from the design stage into all the phases of a product’s lifecycle, we propose the use of: • LiMMA: Lightweight Models with Multilayered Annotations Combines open, lightweight CAD models with annotations to augment geometric data with supplementary information from PLM phases • RRoRIfE: A registry/repository of Representation Information for Engineering Information Supports decisions relating to migration of file formats for continuous and long-term access and reuse

14. Lightweight CAD Formats (LWFs) • Simpler formats: do not retain all the richness of a native CAD model • Analogous to the notion of desiccated formats • Reduced file sizes via compression techniques • Platform and application independent • Easier to curate and preserve • Simpler and more open specifications • Wider and more affordable software support • Smaller files sizessupport distributed exchange • Only contain as much information as a particular recipient needs • Variety of LWFsin current use: • Universal 3D, 3D XML, JT Format, PLM XML, PRC, X3D, XGL/ZGL • Each with properties better suited to some purposes than others • Suitable for access, collaboration and dissemination as well as preservation • Alleviate collaborative exchange, offer customised views, security of IP, application independence, reduced file sizes etc.

15. Multilayered Annotation • Need feedbackfrom downstream processes • Augment geometric model with information from different phases of PLM • Annotation allows incorporation of varied information: • Design rationale, in-service data and other information useful in an iterative design process e.g. context, provenance etc. • Extra information needed for a certain viewpoint • Embedding of security levels to restrict access to certain partners or users • “stand-off” (vs. “in-line”) annotation • Stored separately to CAD model; actual model unchanged • Can be selective about which annotations are shared • Same annotation can be applied to derivatives of full CAD model • Use ofXML for machine-processing

16. LiMMA LiMMA (Lightweight Models with Multilayered Annotations) • A framework for representing CAD models using LWFs with additional layers of XML-encoded Annotations • Annotation layers for • different viewpoints e.g. design, manufacture, service, marketing • different security and access levels

17. LiMMA LiMMA (Lightweight Models with Multilayered Annotations) 1. Designers create several annotation documents, at the same time as the CAD model.

18. LiMMA LiMMA (Lightweight Models with Multilayered Annotations) 2. CAD model can be derived to different LWFs for different users requiring different viewpoints: e.g. PRC for manufacturing engineers; JT for assembly engineers

19. LiMMA LiMMA (Lightweight Models with Multilayered Annotations) 3. All users throughout the product lifecycle are able to feedback their knowledge and experiences through annotations rather than modifying the CAD model itself. e.g. service staff can provide information regarding wear and tear.

20. Annotation Environments Annotation interfaces in UGS NX CAD package and 3D PDF viewer Adobe Acrobat

21. OAIS Representation Information • A wide range of LWFs with differing characteristics are available • Need format appropriate not only for a particular use and view of the product, but also for long-term retention • Open Archival Information System (OAIS) Reference Model • Preservation and curation environment for all types of data • Developed by CCSDS, published as ISO14721:2003 • Defines a functional model and an information model • Representation Informationis any information required to render, interpret, process and understand data • e.g. Software tools for rendering and interpretation; file format specifications and references; conventions and semantics used • Essential for both curation and contemporary use • A planning tool - Registry/Repository of Representation Information for Engineering (RRoRIfE)

22. RRoRIfE (1) Registry/Repository of Representation Information for Engineering • Holds two XML schemata (derived from engineering use cases) relating to: • significant properties of CAD file formats e.g. 2D and 3D geometric entities; modelling history; size; compression; geometry-related metadata (e.g. tolerances, finishes) • characteristics of processing software; record of how well each property is preserved for each format conversion –None; Good; Fair; Poor • Ball, Patel & Ding paper –iPRES 2008

23. RRoRIfE (2) Facilitates simplepreservation planning queries: • Characteristics of file formats and conversion tools • Select a characteristic and see which format supports it • Generate possible migration pathways between two formats • Given a start format and significant characteristics, generate a list of suitable destination formats and conversion pathways

24. Further Work & Conclusions • Further Work • Develop a more integrated framework • Persistent identification of geometry between native CAD formats and the LWFs derived from them • RRoRIfE: accumulate a corpus of information on significant properties of file formats and characteristics of conversion tools • Non-intrusive and automated capture of information in PLM • Selection and appraisal of product lifecycle data • Conclusions • Digital curation challenges posed by PLM are significant • Several tools and techniques specific to improving the robustness and sustainability of CAD file formats • LiMMA: Lightweight Models with Multilayered Annotations • RRoRIfE: Registry/Repository of Representation Information for Engineering information

25. Questions? Thank you Manjula Patel UKOLN, University of Bath, UK m.patel@ukoln.ac.uk www.ukoln.ac.uk Knowledge and Information Management Through Life (Grand-Challenge) Project Digital Curation Centre Innovative Design & Manufacturing Research Centre UKOLN, University of Bath