4. Cross-functional change review board
A restraint design change rarely stays inside one discipline. Moving a seatbelt anchor 15 mm to improve comfort can shift the D-ring routing, alter the CAE belt-pull load path, invalidate a sled test, change the weld sequence on the body-in-white line, and trigger a new PPAP submission from the webbing supplier. The Cross-Functional Change Review Board (CRB) exists to catch those ripple effects before they reach production. It is not a rubber-stamp committee; it is the decision forum where each function contributes the evidence needed to accept, reject, or conditionally release a change.
This chapter defines what each function owns inside the CRB, what inputs it must bring, and how the board converts those inputs into a disposition. The focus is on restraint-specific practice: airbags, seatbelts, pretensioners, load limiters, anchors, and the occupant sensors that trigger them.
Why a Dedicated Board for Restraints
Restraint systems are safety-critical and regulated. A change that looks cosmetic can alter occupant injury metrics (HIC, chest deflection, femur load) measured against FMVSS 208, ECE R94/R95, or C-NCAP protocols. Because the system is validated as a whole, no single function can approve a change alone. The CRB provides a single traceable decision record that links the engineering change request (ECR) to the evidence that justifies it.
Typical triggers for convening the board include:
- Geometry change to an anchor, retractor, buckle, or airbag module mounting point.
- Material or gauge change in a structural component that carries restraint loads.
- Supplier process change (new tool, new plant, new sub-tier source).
- Software or calibration change to the airbag control unit (ACU) or occupant classification system.
- Field or warranty issue requiring a corrective change.
Function-by-Function Responsibilities
The table below summarizes what each member owns. The sections that follow explain the evidence each function must actually produce.
| Function | Primary Responsibility in CRB | Key Evidence / Input |
|---|---|---|
| Design (Restraint Engineering) | Owns the change definition and CAD/geometry impact | Updated 3D model, 2D drawings, tolerance stack, interface matrix |
| CAE | Predicts occupant and structural response | LS-DYNA sled/full-vehicle model results, injury criteria deltas |
| Test / Validation | Confirms physical performance and regulatory compliance | Sled test matrix, dummy instrumentation, pass/fail vs. targets |
| Manufacturing | Assesses process feasibility and assembly impact | Process FMEA update, fixture/tooling change list, cycle time |
| Quality | Guards conformity, traceability, and PPAP status | Control plan delta, inspection strategy, PPAP level |
| Suppliers | Declare feasibility, cost, and timing for their scope | Feasibility commitment, tooling lead time, sample submission date |
Design: Defining the Change Precisely
Design opens the review by stating the change in measurable terms. "Improve belt comfort" is not a change definition; "relocate the B-pillar D-ring 15 mm rearward and 8 mm upward, part number 8R0-857-XXX revision C to D" is. Design must present the revised CAD, the affected drawing zones, and a tolerance stack showing whether the new position still satisfies the belt geometry envelope. For restraint work, the critical check is the belt routing angle at the D-ring and the resulting retractor pull direction, because these directly affect spooling and lock-up behavior.
CAE: Predicting the Occupant Response
CAE translates the geometry change into predicted injury metrics. For a D-ring move, the analyst updates the belt path in the sled model, re-runs the frontal 56 km/h pulse, and reports the delta in HIC15, chest deflection, and belt force. A representative result set might show HIC15 rising from 620 to 655 (still below the 700 target) while chest deflection stays flat. CAE must state the model version, correlation status against the last physical test, and any assumptions. If the model is not correlated for the affected load case, CAE must flag that the prediction carries higher uncertainty and that physical testing is mandatory.
Test: Confirming with Hardware
Test owns the physical proof. Based on the CAE delta, the validation engineer proposes a test matrix: which pulse, which dummy (Hybrid III 50th percentile, THOR, or a child restraint), which seat position, and how many repeats. For a minor geometry change with a small predicted delta, a single confirmation sled may suffice; for a change near a regulatory threshold, the full matrix is required. Test reports raw channels (head acceleration, chest potentiometer, belt load cell) and the pass/fail judgment against the program targets, not just a summary number.
Manufacturing: Feasibility and Process Impact
Manufacturing checks whether the change can be built at rate. A D-ring move may require a new weld nut location, a revised fixture, or a change to the operator's torque sequence. Manufacturing updates the Process FMEA, lists tooling and fixture changes with lead times, and confirms the change does not break the assembly sequence. If the new position forces a re-orientation of the retractor during installation, that is a cycle-time and ergonomics issue that must be resolved before release.
Quality: Conformity and Traceability
Quality ensures the change is controlled from a conformity standpoint. This means updating the control plan, defining the inspection method for the new geometry, and determining the PPAP level required from affected suppliers. Quality also verifies that the change is traceable: the ECR number, the affected VIN range, and the effective date must be recorded so that any field issue can be traced back to the correct build configuration.
Suppliers: Commitment and Timing
Suppliers declare feasibility for their scope. A webbing or retractor supplier must confirm whether the new routing stays within the validated spooling envelope, whether new tooling is needed, and what the sample submission date will be. Suppliers also provide cost impact and any sub-tier changes. The CRB cannot release a change without a supplier feasibility commitment, because an unconfirmed supplier is an unconfirmed change.
How the Board Reaches a Disposition
The board does not vote by seniority; it checks whether the evidence package is complete. A change is released only when all six functions have submitted their inputs and no function has an open red flag. If CAE predicts a metric above target, or Test cannot confirm within the available window, or a supplier cannot commit to timing, the disposition is "hold" and the ECR returns to the originator with a specific list of missing evidence.
A practical decision rule used on many programs is the following. The board approves when: (1) CAE delta is within the program margin, (2) the test plan is defined and funded, (3) manufacturing confirms no line-stop risk, (4) quality confirms PPAP level and control plan update, and (5) suppliers confirm feasibility and sample dates. Any single "no" converts the disposition to conditional release, which allows prototype build but blocks production implementation until the gap closes.
Common Failure Modes in Restraint CRB Reviews
- Design-only approval: The change is released on CAD review alone, and the CAE or test impact is discovered after tooling is cut.
- Stale CAE correlation: The model was correlated to an older build; the predicted delta is treated as accurate when it is not.
- Supplier late declaration: A sub-tier process change surfaces after the board has already approved, forcing an emergency re-review.
- Quality traceability gap: The effective VIN range is not recorded, so a field issue cannot be linked to the change.
Each of these is preventable with a disciplined evidence checklist and a clear rule that no function's silence counts as approval. The CRB's value is not in the meeting itself but in the traceable, cross-functional record it produces — the record that proves a restraint change was evaluated by every function that could be affected before it reached a vehicle.