How will model-based definition change the way we perform FAI?

Model-based definition (MBD) does not remove the need for AS9102 First Article Inspection. It changes where characteristics come from, how they are ballooned and flowed down, and how FAI data is captured and traced. In most aerospace and defense environments, MBD will coexist with 2D drawings and legacy systems for many years, so FAI workflows must support both.

What stays the same for FAI with MBD

Even with a fully annotated 3D model and PMI, several fundamentals do not change:

• You still need to comply with AS9102 (or customer-specific FAI requirements).
• You still must identify and account for all design characteristics that affect form, fit, function, and safety.
• You still need traceable measurement results tied to each characteristic and each part/serial.
• You still need documented FAI packages that customers and auditors can review.
• You still need configuration control between the “authority” dataset and the inspected part.

MBD mainly changes how you derive, manage, and consume those characteristics, not the obligation to perform a robust, documented FAI.

What changes with model-based definition

With MBD, the authority dataset is the 3D model with embedded PMI (dimensions, GD&T, notes). That shifts several FAI steps:

1. Characteristic extraction and ballooning

• From manual ballooning on 2D printouts to digital extraction from the 3D model and PMI.
• FAI tools can, in principle, auto-generate a characteristic list (feature ID, nominal, tolerance, GD&T) directly from the model.
• Ballooning becomes feature- or PMI-based in 3D instead of 2D callouts.

However, this only works reliably if:

• PMI is complete and unambiguous.
• There is clear designation of which dataset is the authority (customer vs. internal model, model vs. drawing).
• Your FAI/inspection software correctly interprets the CAD format and PMI conventions.

Where PMI is incomplete or inconsistent, you will still need manual review, supplemental 2D drawings, and engineering clarification, just as in drawing-based FAI.

2. Planning and linking inspection to features

• FAI planning can be generated directly from features on the model, enabling better linkage between:
• CAD / PLM (design intent)
• CAM / CNC (how it is machined)
• CMM / inspection programs (how it is measured)
• Characteristic IDs can be consistent across systems (PLM, MES, QMS, CMM) if you implement good ID governance.

This can reduce transcription errors and rework, but only if:

• Your PLM, CAM, CMM and FAI tools are well integrated and validated.
• Feature and characteristic IDs are under change control and not constantly renumbered by design changes.
• There is a clear mapping for legacy parts that still rely on 2D ballooning.

3. CMM and automated measurement integration

• MBD can feed model-based CMM programming, allowing inspection paths and features to be auto-derived from the same 3D model used for FAI.
• Measured values can be pushed back automatically into an FAI form or database by characteristic ID.

This reduces manual data entry, but introduces new risks and requirements:

• CMM software must reliably read the exact CAD/PMI version used for manufacturing.
• Automated data transfer must be validated to avoid silent mis-mappings (wrong characteristic linked to the wrong result).
• Changes to CAD models, CMM programs, or FAI mappings must go through change control with impact assessment on past and in-process FAIs.

4. FAI forms and digital evidence

• Instead of manually populating an AS9102 form from 2D balloons, you can have a digital FAI record linked directly to the model, PMI, and measurements.
• System-generated FAI reports (including AS9102-like layouts) can be exported from an FAI/QMS/MES system drawing directly on the characteristic database.

However, most customers and auditors will still expect:

• Human-readable FAI evidence (PDF reports, 2D views, or snapshots of the model with clearly identified characteristics).
• Ability to review the FAI without your specific CAD/PLM tools.

Many organizations end up with a hybrid: model-based planning and data capture internally, and exported AS9102 forms with 2D callouts or screenshots externally.

5. Traceability and configuration management

MBD can strengthen traceability if implemented correctly:

• FAI records can be tied to specific model revisions and PMI states.
• Downstream NC, deviation, and RCCA records can reference the same characteristic IDs as the FAI.
• Design changes can be compared at the model/PMI level to determine when a partial vs. full FAI is required.

But this depends on:

• A clear configuration management strategy for models, derivatives, and viewables.
• PLM integration with MES/QMS so that the correct model revision is enforced at release and at inspection.
• Governance around when and how to reuse FAI evidence after design updates.

Impacts on process, validation, and training

MBD-driven FAI is not just a CAD upgrade; it is a process and system change that must be deliberately managed.

• Process definition: You will need updated procedures that specify how to perform FAI from an MBD authority dataset, including exceptions and handling of mixed 2D/3D packages.
• System validation: Automated extraction, mapping, and reporting workflows must be verified and, in many cases, formally validated before they can be relied on for compliance evidence.
• Training: Inspectors, manufacturing engineers, and quality engineers will need training on reading PMI, navigating 3D viewers, and recognizing when MBD data is incomplete or suspect.

Brownfield reality: coexistence with 2D and legacy systems

In most aerospace and defense organizations, MBD will roll out by program, platform, or customer. For a long period you will:

• Run FAIs on 2D-only parts, MBD-only parts, and hybrid data sets.
• Maintain multiple CAD formats, PLM systems, or authority definitions across customers.
• Feed FAI results into existing ERP/MES/QMS stacks that were designed around drawings and paper packets.

Because full system replacement is costly and risky in regulated, long-lifecycle environments, MBD-based FAI usually starts as an overlay on top of existing workflows:

• FAI tools that can ingest both 2D and 3D+PMI and output standardized AS9102-like reports.
• Interfaces that push summarized FAI status/results into ERP/MES without requiring replacement of those systems.
• Controlled pilots on new programs before scaling plant-wide.

Trying to replace all legacy systems and processes at once to be “MBD-native” typically runs into qualification burden, validation cost, integration complexity, and downtime risk.

Key tradeoffs to consider

• Efficiency vs. complexity: Automated characteristic extraction and CMM integration can save time, but increase dependence on correct CAD/PMI and robust integrations.
• Standardization vs. flexibility: MBD encourages standardized IDs and workflows, but must accommodate differing customer CAD formats and rules.
• Automation vs. oversight: The more automation you add, the more you need disciplined validation, change control, and monitoring to catch mapping or configuration errors.

Practical steps if you are moving FAI into MBD

• Clarify which datasets are the authority for each customer/program (model, drawing, or both).
• Define how characteristic IDs are created, owned, and changed across PLM, CAM, CMM, and FAI tools.
• Start with one or two representative programs to pilot MBD-based FAI workflows and refine procedures.
• Validate auto-extraction and data mapping, especially where CMM or gage data flows into your FAI records.
• Ensure you can still produce auditor- and customer-friendly AS9102-style evidence, even if the backbone is model-based.

If you treat MBD as a new authority for design and a better source of FAI characteristics, rather than a shortcut around AS9102 requirements, it can significantly improve traceability, reduce manual effort, and tie FAI more tightly to your digital thread.