Capture and Reuse of Engineering Knowledge in Digital Twins

Today, industry is struggling to integrate data from different software applications and domains. Standards based solutions, such as ISO 10303, are increasingly recognized to solve this issue. A consistent and comprehensive digital twin may serve as a knowledge base of a product in operation. Design intent may be included as well as experiences throughout the life of the product. Today's lack of access to such information is a major hinder for innovation. Great effort is spent today on regaining knowledge that once existed. These are the issues that this PhD project will address. It, thus, touches the strategic importance of engineering data as a representation of knowledge for industry. The project has after its first year concluded how to apply and enhance the ISO 10303-209 (AP209) data model to represent structural test data consistently with related simulation data. This model will enable data management across design, finite element method (FEM) analysis and structural testing. The following new types of data can now be represented by AP209: - Properties of sensors used in physical tests - Tests performed on a product - Relations between sensors and FEM elements - Relations between tests and sensors - Relations between tests and FEM load cases - Relations between structural test results, sensors and tests - Positions and orientations of sensors The project has also integrated analysis and structural test data with the PDM and configuration meta data approach typically used for engineering design and has applied this in a SDM tool. [Year 2] In the 2nd year of the project a complex aircraft part was designed, analysed and manufactured, and later structurally tested. - The work regarding the application and enhancement of the STEP AP209 standard was continued, specially by implementations in the SDM tool. - A converter for processing structural test and FEA data created during the first year was greatly improved during the second year. - A tool processing AP209 models containing structural test and FEA data was developed. The application allows to investigate structural test results, FEA results and their relations using graphical plots. - A tool processing AP209 models containing sensors and FEA models was created and allows the display of strain gages on their related FEA mesh. - Completed and published the paper "Relating structural test and FEA data with STEP AP209" in "Advances in Engineering Software", As STEP is a very advanced format, it can be complicated to validate STEP files/models created by different tools. For this reason, a tool named STEP Explorer was developed. The tool is used to visualize and explore the content of STEP models in a very simple and efficient way. It is meant for better understanding the STEP standard, checking custom STEP models, creating graphical diagrams etc. New research started on a related topic; how to add nonlinear analysis to AP209: - STEP AP209 was designed to cover linear static and linear modal analyses. An investigation on how the standard should be improved to cover more advanced analyses, such as nonlinear analyses was started. - Investigated different well-known simulation tools and looked at how the analyses they cover could be standardized in STEP AP209 - Started developmen of a converter between Ansys and AP209 - A paper has been started on this topic [Year 3] In the third year of the project, the research on how to enable AP209 to cover nonlinear analysis was continued. An extended version of AP209 that covers the following FEA topics was made: - Classification of analysis types based on if the analysis is; static or dynamic, buckling, linear or nonlinear, modal or frequency response, implicit or explicit, etc. - Generic approach of how to represent analysis parameters (f.ex. related to time increments, iterations, convergence, etc.) - Time and space varying loads and boundary conditions - Element contact - Nonlinear materials The converter which was started in year 2 was improved and covers the following conversions: - NASTRAN to and from AP209 - AP209 to ANSYS - Abaqus to ANSYS FEM test cases were introduced covering the main content of the new AP209 model. These were modeled in Abaqus, and using the converter, translated to Nastran and Ansys. By solving and comparing the results of each test cases, the new AP209 model could be validated by showing that all necessary data was covered by it. This work was described and presented in two academic papers: 1) "ISO 10303 AP209 - Why and how to embed nonlinear FEA" (under review) 2) "Extending ISO 10303-209 for Nonlinear Finite Element Analysis" (submitted) Both papers were sent to the journal Advances in Engineering Knowledge; one is under review, the other is waiting for review. A third paper, covering FEM contact and nonlinear materials, was also planned, but was instead included as an appendix to the PhD thesis.

The ISO 10303 AP209 standard is not well known by all who could take advantage from it. Three academic papers related to AP209 have been produced, which will foster scientific discussion of data interoperability issues and helped to start several R&D projects; these will make the standard more available for the industry. Results of this PhD project are expected to be documented in updated versions of the standard itself. AP209 will enable industry and software vendors to increase the efficiency of their processes and those of their customers. This may lead to more effective data interoperability, data management and long term archiving of simulation and test data. During an on-going industry project, additional use cases involving data that are not covered in this thesis, will be addressed. This and future projects will require additional AP209 recommendations and extensions, but will be based on the current results of this PhD project.

The Aerospace and Defense industry addresses today configuration control of product data and their sharing by frameworks for data management and for data interoperability. Data management is resolved by a holistic approach that collects all product related data or references to them in one location. This leads to a comprehensive and consistent repository of product lifecycle data, a digital twin. Applications for product lifecycle management, PLM, and simulation data management, SDM, exist, but they do not provide the required sustainable architecture as they are not openly interoperable and standards based. They do not give control to industry over their own data. Interoperability of product data is resolved by applying standards for the exchange and sharing of data, first of all ISO 10303, STEP. The scope of engineering data is best supported by ISO 10303-209, Multidisciplinary analysis and design. This project will benefit manufacturers and users of complex engineering assets, such as, aircraft, ships and oil&gas installations. They need to be given full control over their product data and associated knowledge throughout the potentially long lifetime of the physical twins. Long term data retention of engineering data and knowledge need to be resolved to allow manufacturers to produce and deliver comprehensive sets of design, simulation and test data to enable user organizations to support and upgrade such assets. Jotne is cooperating in R&D projects with relevant manufacturers, such as, Airbus, Boeing and Lockheed Martin. This PhD study will interact with these companies to collect requirements, receive test data and evaluate solutions. The project will also communicate closely with the ISO community for ISO 10303, STEP, to validate and possibly improve digital twin related standards. For Jotne the project will improve our commercial suite of STEP software applications.