3. Regulatory and standards landscape for CAE verification

CAE verification does not happen in a vacuum. Every credible simulation program must be able to answer a deceptively simple question: against what benchmark, and by whose authority, do we declare a model "verified"? The answer comes from a layered landscape of standards — some mandatory by regulation, some contractual, some voluntary but expected by customers and regulators alike. This chapter maps that landscape and, more importantly, shows how to translate each standard into concrete CAE program requirements.

The Four Pillars

For most engineering organizations, the standards landscape rests on four pillars:

  • ASME V&V 10 / 20 / 40 — the de facto methodology standard for computational solid mechanics (CSM) and computational fluid dynamics (CFD). ASME V&V 10 defines verification and validation terminology and procedures; V&V 20 covers CFD; V&V 40 covers CSM. These are methodology standards: they tell you how to verify and validate, not what acceptance criteria to use.
  • ISO standards — ISO 9001 provides the quality management framework; ISO/IEC 17025 governs testing and calibration laboratory competence; ISO 16750 and ISO 26262 (now superseded by ISO 21448 for SOTIF) address environmental and functional safety requirements for automotive electronics. ISO standards are typically process standards: they require documented, repeatable, traceable procedures.
  • NASA-STD-7009A — the NASA Standard for Models and Simulations. This is the most prescriptive of the four, requiring a formal Credibility Assessment Scale (CAS) across eight factors: verification, validation, input pedigree, uncertainty characterization, results robustness, use history, M&S management, and people qualifications.
  • Industry-specific requirements — nuclear (ASME NQA-1, 10 CFR 50 Appendix B), aerospace (DO-178C for software, DO-254 for hardware), medical devices (FDA 21 CFR Part 820, ISO 13485), and automotive (IATF 16949, VDA 6.3). These add domain-specific evidence and documentation requirements on top of the methodology standards.

How the Standards Interact

The pillars are not mutually exclusive. A typical aerospace CAE program might use ASME V&V 40 for the verification methodology, NASA-STD-7009A for credibility assessment, ISO 9001 for the quality management system, and DO-178C for any simulation software that is part of a certified system. The key insight is that methodology standards define the "how," process standards define the "who and when," and domain standards define the "what evidence."

Standards Landscape for CAE Verification ASME V&V Methodology Standard • V&V 10: Terminology • V&V 20: CFD • V&V 40: CSM Defines: verification procedures, validation metrics, uncertainty ISO Standards Process Standard • ISO 9001: QMS • ISO/IEC 17025: Lab • ISO 16750: Automotive Defines: documented, repeatable, traceable procedures NASA-STD-7009A Credibility Standard • 8-factor CAS • Verification evidence • Input pedigree Defines: credibility assessment scale, evidence requirements Industry-Specific Domain Standard • Nuclear: NQA-1 • Aerospace: DO-178C • Medical: 21 CFR 820 Defines: specific evidence, audit trails, certification paths CAE Verification Program Evidence · Traceability · Credibility How to verify Who & when What evidence Domain-specific proof

Translating Standards into Program Requirements

Reading a standard is one thing; turning it into a working CAE verification program is another. The practical translation follows a three-step pattern:

  1. Map the standard's clauses to your simulation lifecycle. For ASME V&V 40, Clause 3 (Verification) maps to your mesh convergence studies and code verification; Clause 4 (Validation) maps to your physical test correlation activities. For NASA-STD-7009A, each of the eight CAS factors maps to a specific evidence artifact — for example, "input pedigree" maps to a documented source and uncertainty for every material property and boundary condition.
  2. Define acceptance criteria that satisfy the strictest applicable standard. If your program must satisfy both ASME V&V 20 and a customer's internal CFD guideline, use the more demanding criterion. A common example: ASME V&V 20 recommends a Grid Convergence Index (GCI) below 5% for integral quantities; a customer may require below 3%. Your program should default to 3%.
  3. Build traceability from requirement to evidence. Every verification activity must be traceable to a specific standard clause, and every piece of evidence must be traceable to a specific model version, input deck, and solver version. This is where ISO 9001's document control requirements become concrete: your CAE file management system must enforce versioning, access control, and audit trails.

A Practical Example: CFD Verification for a Heat Exchanger

Consider a CFD verification task for a shell-and-tube heat exchanger, governed by ASME V&V 20 and an internal quality procedure aligned with ISO 9001. The verification plan would include:

Activity Standard Reference Acceptance Criterion Evidence Artifact
Code verification (method of manufactured solutions) ASME V&V 20, Section 3 Observed order of accuracy ≥ 1.8 (2nd-order scheme) MMS report with grid refinement table
Grid convergence study ASME V&V 20, Section 4 GCI ≤ 3% for outlet temperature Three-grid Richardson extrapolation
Validation against experimental data ASME V&V 20, Section 5 E ≤ 5% for overall heat transfer coefficient Comparison plot with uncertainty bands
Input uncertainty documentation NASA-STD-7009A, Factor 3 All inputs have documented ± range Input pedigree table
Solution verification (iterative convergence) ASME V&V 20, Section 3 Residuals < 1e-5, monitors stable Solver log and monitor history

Note that the acceptance criteria are example values chosen for illustration; actual values must be derived from the specific application, fluid properties, and required accuracy. The key point is that each row links a standard clause to a measurable criterion and a concrete artifact.

Practical tip: When multiple standards apply, create a "standards matrix" that lists each standard, its applicable clauses, and the corresponding program activity. This matrix becomes the backbone of your verification plan and the first document an auditor will request.

Common Pitfalls and How to Avoid Them

  • Treating standards as checkboxes. A program that claims ASME V&V 40 compliance but cannot produce a grid convergence study for a critical model is not compliant. Standards require evidence, not declarations.
  • Ignoring the hierarchy. Domain-specific standards (e.g., 10 CFR 50 Appendix B for nuclear) often impose stricter documentation and QA requirements than the methodology standards. Always check the domain layer first.
  • Confusing verification with validation. ASME V&V 10 is explicit: verification is "the process of determining that a computational model accurately represents the underlying mathematical model"; validation is "the process of determining the degree to which a model is an accurate representation of the real world." Your program must keep these activities separate and document both.
  • Neglecting software quality. If your CAE solver is part of a certified system (e.g., DO-178C for airborne software), the solver itself must be verified to the appropriate software level. This is a common gap in CAE programs that focus only on the physics.

Building the Standards Foundation

The regulatory and standards landscape is not a constraint to be endured — it is the foundation of a defensible CAE verification program. By mapping each applicable standard to specific activities, acceptance criteria, and evidence artifacts, you transform abstract requirements into a repeatable, auditable workflow. The next chapter will build on this foundation by showing how to construct a verification plan that satisfies multiple standards simultaneously without duplicating effort.

← Previous 📖 Contents Next →