Bridging the Validation Gap: Next-Level Model Governance in Digital-Twin Engineering
0 likes23 views
The document discusses the need for enhanced model governance in digital-twin engineering to address complexity and ensure model quality in systems engineering. It outlines the challenges faced, including technical debt and compliance issues, while presenting Incquery's solutions for systematic model governance and quality assurance. The aim is to improve collaboration and efficiency across projects while maintaining standards and reducing errors.
Bridging the Validation Gap: Next-Level Model Governance in Digital-Twin Engineering
- 1. Denver, CO, USA Vienna, Austria Budapest, Hungary Bridging the Validation Gap: Next-Level Model Governance in Digital-Twin Engineering A Journey to Consistency Speaker Dr. Ákos Horváth, CTO IncQuery Group
- 3. Co-creating digital engineering success We can help with • Complexity challenges • Boosting MBSE buy-in • Increasing productivity We are what you need • Full service partner • Co-processing developer • Business innovator We deliver what you demand • Consultancy by experts • End-to-end toolchain architecture • Tailor-made engineering tools • Excellence within budget The IncQuery Group is an international team of engineering experts with a strong R&D background. We strive to be the trusted industry partner for engineering organizations working in Mission- and Safety Critical contexts, to get their work done easier and faster.
- 4. Context MBSE Digital Thread, Digital Data Fabric, Digital Engineering System of System Engineering, AI-Driven Digital Engineering, … and the list goes on
- 5. From • Constructing buildings, • Aqueducts • Bridges • Ships • Sewers • Steam locomotives To • Very large systems • Systems of systems • Complex socio-technical systems • Smart * • and in situ medical devices Copyright Incose 2035 vision: http://sevisionweb.incose.org/ The past - System Engineering of the old
- 6. Adaptation of SE • varies across industry domains, organizations, and projects. Practices • Well-founded in • Requirement management • Functional analysis • Safety assessment • Emerging • Knowledge representation • Multi-domain sim and processing Complexity growth is exponential • Software Complexity (over 1BN Loc for ADAS) Tool landscape • Specialized tools for each discipline • Proprietary data formats • Limited standardization Source: INCOSE 2035 vision: http://sevisionweb.incose.org The present - Systems Engineering of today
- 7. • Wider range of standardization and enterprise adaptation • Integrate other disciplines such as data science, hybrid simulation • Formal systems engineering theoretical foundations • AI will both impact • the systems engineering practice • and the systems to be designed Source: INCOSE 2035 vision: http://sevisionweb.incose.org The future – Challenges of System Engineering
- 8. One possible solution approach – Digital-twin pipelines Source: “Taming Monsters With Dragons or How to Build Digital Twins The Right Way” - Robert Karban
- 9. What is “engineering work” quality? Model quality and technical debt
- 10. What Is Model Quality? Free from Common Mistakes Methodology Conformance Standards Compliance
- 11. Example: “Common Mistakes” Basic SysML omission error: State Machine has an outbound Transition from a Choice Pseudostate that does not have a Guard Impact: “Accidental” ambiguity → Model interpretation may be not be consistent across the team → Unintended behavior may appear in the system → Potential cybersecurity issue in the end product https://www.saic.com/digital-engineering-validation-tool
- 12. Example: Standards Compliance Common SysML Compliance Issue Proxy ports must be typed by an Interface Block (SysML v1.6 Section 9.3.2.13 Constraints) Impact Chain: We think we are using SysML correctly as specified → Our model may be incompatible with other SysML tools → Simulations may fail and other downstream tools may report errors → Rework → Missed deadlines → Budget overruns
- 13. Example: Methodology Conformance Common Gap: Missing Trace Relationships Each Use Case element must have «refine» or «trace» relationships to a Requirement element Impact Chain: Non-conformance to MBSE method traceability rules → Missing traceability link → ISO26262 violation → Certification audit finding → Rework → Missed deadlines → Budget overruns
- 14. What Are The Sources Of Model Quality Issues? Lack of Expertise Invisible Errors Awareness of Readiness Modeling Tool != Drawing Tool
- 15. Technical (/Complexity) Debt • Represents the deferred costs of repairing, reworking, or replacing a product that wasn’t built correctly from the beginning o Like financial debt, it can be a useful tool when used at the right time for the right purposes o The goal is not to eliminate debt but to make better informed choices about debt • Model rework is often much more difficult than for software o More often architectural in scope and nature o Implementation details are exposed to users o Refactoring of models is less doable piecemeal o Automated discovery is more difficult o Defect “contagion” is common "The Nature of Technical Debt in the Development of Descriptive Models in MBSE" Ryan Noguchi, Aerospace Corp., INCOSE IS 2024
- 16. How can we help? IncQuery’s approach for Systematic Model Governance
- 17. Quality criteria that define the preconditions for data exchange in engineering workflows. Enforce policies (style guidelines) across several stakeholders (organizations). Make sure that models are • Consistent in style (standards conformance); • Understandable across several teams (methodology conformance); • Free of common mistakes & anti-patterns; • “Complete”. Model Governance
- 18. Quality criteria that define the preconditions for data exchange in engineering workflows. Enforce policies (style guidelines) across several stakeholders (organizations). Make sure that models are • Consistent in style (standards conformance); • Understandable across several teams (methodology conformance); • Free of common mistakes & anti-patterns; • “Complete”. Model Governance
- 19. Standalone • Add-in for Desktop • Incremental rule execution → up to 10X faster than built in solution Team • DevOps pipeline deployment, integrated with Version Control • Rich report documents with Quality Score Enterprise • Execute model governance at scale • Aggregated quality metrics • Historical updates • Dashboard-based reporting Configurable rule set execution at the required place and time. IncQuery Validator for an “Ecosystem”
- 20. SEs Collaborative work Standalone - IncQuery Validator for Helps SE to efficiently, execute modelling activities and deliver error-free models. Individual responsibility, process-driven.
- 21. SEs QA Collaborative work Governance Automation Review Audit Reporting Teams - IncQuery Validator for Helps Teams to collaborate efficiently, by maintaining consistency and high quality across multiple projects.
- 22. Enterprise - IncQuery Validator for Project 1 Project 2 IncQuery Validator: extract data & evaluate model quality IncQuery EKL: Database to store data and metadata (historic results) File (JSON/CSV) File share (Sharepoint) Document DB Graph DB API Displays data in dashboards DevOps pipeline On-prem / Cloud DB cache publish extract store Review Audit Trends Validation Reporting Dashboards SEs QA Helps the whole Company to collaborate efficiently, by maintaining consistency and high quality across all projects and all stakeholders
- 23. Dashboard Overview – Current status + history
- 24. Dashboard Overview – Common issues
- 25. Dashboard Overview - Blame
- 26. Summary - today • SE paradigm shift to formal languages and automation, i.e. MBSE • Digital threads • Quality assurance • Automation • Systematic qualification and audit trail • Close gap between engineering documents and models • Standards based and commoditized
- 27. Summary - future On the market R&D&I - next steps IncQuery Validator for SysMLv2 Synchronized System wide simulation & System Engineering GNSS + Autonomous driving validation using graph solvers (Refinery) Pragmatic formal verification of SysML (v2) Environment for model-based rigorous adaptive co-design and operation of CPS
- 28. Thank you! akos.horvath@incquery.io US +1 720 655 6118 +1 720 655 8285 EU + 36 30 774 2839 Looking forward to a long-lasting partnership! Visit our website Book a Demo with Our Tech Experts Access Our Case Studies Subscribe to our newsletter