MBSE Panel Integrating MBSE into a MultiDisciplinary Engineering
MBSE Panel Integrating MBSE into a Multi-Disciplinary Engineering Environment A Concurrent Engineering Perspective Hans Peter de Koning (European Space Agency / ESTEC, Noordwijk, The Netherlands) 20 June 2011 MBSE Panel | INCOSE IS 2011 | Denver, CO, USA 1
My background Concurrent Design Facility (CDF) at ESA/ESTEC • Started in 1998 • New advanced facility in 2008 • 150+ multi-disciplinary studies performed with a model based approach • Typical study is 8 half-day sessions in 4 weeks and noncolocated work in between • Typically 20 disciplines in one team Main activity: • Conceptual design of ESA space missions (Phase 0) Also: • Major project reviews • Anomaly investigation boards For details, see www. esa. int/cdf MBSE Panel | INCOSE IS 2011 | Denver, CO, USA 2
Illustration of CDF Design Process • Parametric design using a spiral, iterative approach • Each discipline is responsible for a number of parameters (output) • A discipline can subscribe to any parameter of other disciplines (input) MBSE Panel | INCOSE IS 2011 | Denver, CO, USA 3
Q 1 – MBSE expectations Systems engineering / Other engineering disciplines (1/2) Ø Main problem today is integration of models of all disciplines – All individual disciplines have mature modeling methods and tools – except systems engineering ; -) – However, mostly poorly connected islands → Discipline & model integration is where the next major efficiency gain can be made Ø MBSE repository: a natural multi-disciplinary "model/data hub" – Orchestrate overall analysis, design, modeling and V&V – Provide repository for common data that all disciplines need § Requirements, functional decomposition, logical and physical architecture, interface def's, V&V def's, traceability linking all – Ensure consistency among individual discipline models – Manage, aggregate and monitor main budgets, KPIs, Mo. Es, … – Enable multi-disciplinary trades and optimization MBSE Panel | INCOSE IS 2011 | Denver, CO, USA 4
Q 1 – MBSE expectations Systems engineering / Other engineering disciplines (2/2) Ø Other disciplines bring established methods and tools – W. r. t to "model-based" they are more advanced and mature than systems engineering, e. g. : M-CAD, FEM, CFD, E-CAD, Control Simulators, Logistics Simulators, CASE/MBSWE, … Ø MBSE is now going through the transition E-CAD and M-CAD made from mid 70 -s to mid 90 -s and CASE/MBSWE in the last two decades MBSE Panel | INCOSE IS 2011 | Denver, CO, USA 5
Q 2 – What can MBSE learn from model-based approaches used in other engineering disciplines? (1/2) Ø M-CAx / E-CAx - design, analysis, simulation, V&V – Geometry and topology – 2 D, 3 D, CSG, parametric design – Structured modular approach (separate part and assembly definition) – Electronic catalogues / libraries of reusable items / materials – Integrated configuration / variant / revision management (PLM) – Integrated digital approval / sign-off (PLM) – Scaleability to big models, including analysis/simulation results Ø Control engineering and system simulation – Executable block diagrams – Modeling languages: § MATLAB/Simulink and derivatives § Modelica and similar object oriented simulation languages MBSE Panel | INCOSE IS 2011 | Denver, CO, USA 6
Q 2 – What can MBSE learn from model-based approaches used in other engineering disciplines? (2/2) Ø Software engineering (and IT) – Object-oriented models – separation of concerns, inheritance, encapsulation, polymorfism, … – GUIs and graphical notation relevant to system engineering – e. g. UML → Sys. ML – Configuration / version control – "trunk", branching, merging – Time-boxed, incremental, iterative "agile" development approach as opposed to traditional waterfall → has consequences for procurement / contractual approach – Test-driven SW development → V&V-driven system engineering? – Wikis / multi-authoring → useful to bridge documents / models Ø Data modeling and database design – Conceptual data modeling / ontologies / semantic web – formal rigour, precision and automated reasoning – Concurrent multi-user environments – RDBMSs, no. SQL DBs, ACID principles, scaleability, replication, distribution MBSE Panel | INCOSE IS 2011 | Denver, CO, USA 7
Q 3 – How should the practices and tools be integrated/coupled across disciplines? (1/2) Ø Federation of databases with a central SE hub Ø Need common formal semantic / conceptual data model – To enable consistent mapping / conversion of data – To define an end goal / ideal architecture to work to Ø Loose coupling to allow for adaptation to legacy tools Ø Keep pushing tool vendors for open interfaces Ø Evolutionary migration path, taking small steps at a time – Try to find champions that want to start with an MBE approach – Plan for some management reserve § To provide document deliverables for contractual reasons § To assist the transition and train engineers MBSE Panel | INCOSE IS 2011 | Denver, CO, USA 8
Q 3 – How should the practices and tools be integrated/coupled across disciplines? (1/2) Ø Do not standardize on tools but create open standards (protocols and APIs) for connecting tools and databases – Ideally supported by free open source middleware layers – maintained by a vendor-neutral community of practioners / standardization body – Invest in automated model transformation and code generation to keep complex IT infrastructures manageable and affordable E. g. approach taken in ECSS E-TM-10 -23 "Space system data repository" MBSE Panel | INCOSE IS 2011 | Denver, CO, USA 9
Q 4 – How are the system, hardware, and software models managed to ensure an integrated technical baseline? Ø Learn from PLM – Integrated digital approval / sign-off – Sometimes too tightly integrated with CAx tools – therefore difficult / expensive to extend outside original toolset Ø Learn from open source software – Collaboration portals: Sourceforge, Github, Atlassian, Google Code, Code-Beamer, … – Configuration control: Subversion, GIT, Mercurial, … Ø Keep configuration / version management orthogonal to engineering model content and modeling tools Ø Model (and results data) management will remain a hot topic – No easy, simple solutions – Even more so in concurrent multi-disciplinary environments MBSE Panel | INCOSE IS 2011 | Denver, CO, USA 10
Q 5 – How should a program be organized to achieve more effective utilization and application of MBE? Ø Allow / require models as contractual deliverables – Promote transition from documents to models – keeping (generated) documents for the time being as formal deliverables – Provide model viewing tools with mark-up capabilities to support review and shadow engineering process Ø Go for more iterative, incremental approach – Example from SW engineering: time-boxed "agile" process Ø Develop and give training courses – Including worked examples and "getting started" guides Ø Library of "good" modeling examples and patterns (templates) Ø Try to find some key problems that were solved with MBE solutions – build convincing MBE stories – Hard ROI will always be very difficult – Well-documented stories can be convincing MBSE Panel | INCOSE IS 2011 | Denver, CO, USA 11
- Slides: 11