4. Change impact assessment workflow

In a crash simulation model lifecycle, design changes arrive continuously: a thicker B-pillar, a relocated seat anchor, a new material grade for the front rail, a revised bumper beam section. Not every change invalidates the validated model, and not every change can be absorbed by a quick mesh update. The purpose of the change impact assessment (CIA) workflow is to convert an incoming engineering change into a defensible, traceable decision: does this change require requalification, and if so, at what criticality and scope? Without a structured process, teams oscillate between two failure modes — over-requalifying everything (burning weeks of CAE time on changes that cannot affect the validated response) and under-requalifying (shipping a model whose correlation silently degraded).

The workflow described here is deliberately front-loaded: most of the effort happens before any solver run, in screening and scoping. The output is a signed assessment record that states the change classification, the affected model regions, the requalification scope, the criticality level, and the evidence basis for each conclusion.

Step 1 — Intake and change characterization

Every change enters through a single intake record. The record must capture more than a part number revision; it must describe the change in terms the model can be interrogated with. Minimum fields: change ID and source (design release, ECN, supplier notification, test failure); affected part(s) and their position in the BIW or subsystem hierarchy; the nature of the change (geometry, gauge, material, joint/connection, added/removed part, mass distribution); magnitude (e.g., gauge 1.6 → 1.8 mm, material DP590 → DP780, rail section height +12 mm); and the crash load cases the change is claimed to serve.

A common intake defect is a change described only by its CAD delta. The assessor needs the functional description: a gauge increase on a front rail is a stiffness and mass change in the frontal load path; a seat anchor relocation is a change in restraint attachment geometry and local stiffness that feeds occupant kinematics. Characterizing the change functionally is what makes the downstream screening rules applicable.

Step 2 — Screening against the validated model baseline

Screening answers a narrow question: can this change plausibly alter any response quantity that the model was validated against? The baseline for screening is the correlation matrix from the original requalification — the set of physical tests, load cases, and response channels (e.g., B-pillar intrusion at 40% ODB, rail axial crush force, seat track displacement, HIC36, chest deflection) with their validated tolerances.

Screen each change against three filters:

• Load-path membership. Does the changed part carry load in any validated load case? A bracket on a non-structural closure panel may be outside every load path; a front rail is in all frontal paths.
• Response sensitivity. Is there prior evidence — from the original sensitivity study, DOE, or historical requalification records — that this parameter influences a validated channel beyond its tolerance? Gauge changes on crush members are typically high-sensitivity; trim and attachment detail changes are often low.
• Mass and inertia effect. Does the change shift mass or CG enough to alter global kinematics, occupant loading, or barrier engagement? A 3 kg shift at the rear floor is usually negligible; 15 kg at the front bumper is not.

Changes that fail all three filters are classified No Impact and closed with a documented rationale. Changes passing any filter proceed to scoping. The screening decision must cite the specific validated channel and the specific evidence, not a generic statement.

Step 3 — Scoping the requalification

Scoping determines what must be re-run and re-correlated. The scoping logic works outward from the changed region: identify the affected model zone (parts, connections, contacts), then the load cases whose load paths pass through that zone, then the response channels within those load cases that are sensitive to the zone. The result is a scope matrix — load case × response channel — with each cell marked as re-run required, re-check only (compare against existing results), or unaffected.

Scoping must also address model-build consequences: does the change require re-meshing that alters element quality in the zone, new connections (spot welds, adhesives, bolts), updated material cards, or contact definitions? A geometry change that forces a re-mesh of a validated crush zone is a larger scope than the same change applied to an already-coarse region, because mesh sensitivity itself becomes a requalification concern.

Step 4 — Criticality classification

Criticality sets the rigor of the requalification: the depth of correlation review, the level of independent check, and the sign-off authority. A practical four-level scheme:

Criticality is not the same as change magnitude. A small gauge change on a front rail can be C1 because the channel is regulated and highly sensitive; a large cosmetic geometry change on a non-structural panel can be C4. The classification must be justified by the channel sensitivity and regulatory exposure, not by how large the CAD delta looks.

Step 5 — Assessment record and gate

The workflow closes with a single assessment record containing: change characterization, screening result with cited evidence, scope matrix, criticality level with justification, required actions (re-runs, re-correlations, model-build updates), and the gate decision — proceed to requalification, proceed with conditions, or close as no-impact. The record is reviewed at a change control gate before any solver time is committed. This gate is the cost-control point of the entire workflow: it is far cheaper to challenge a scope decision on paper than to discover mid-requalification that the scope was wrong.

Worked example: front rail gauge change

Consider an ECN raising the front rail gauge from 1.6 mm to 1.8 mm on a DP590 rail, affecting the frontal load path. Intake records the functional description: stiffness and mass increase in the frontal crush zone, +1.4 kg per side. Screening: the rail is in the 100% frontal, 40% ODB, and MPDB load paths (filter 1 pass); historical sensitivity data shows rail gauge drives axial crush force and intrusion by roughly 4–6% per 0.2 mm (filter 2 pass); mass shift is small but front-biased (filter 3 marginal). Scoping: affected zone is the rail, its spot welds to the shotgun and cradle, and the bumper-to-rail connection; load cases are the three frontal cases; channels are rail crush force, toe pan intrusion, and occupant chest deflection in the 40% ODB. Criticality: C1, because chest deflection is a regulated channel and the sensitivity evidence is strong. Required actions: re-run the three frontal cases, re-correlate crush force and intrusion against the validated reference, and run a one-parameter sensitivity check at 1.7 mm to bound the response. The gate approves the scope with the condition that the sensitivity check is completed before correlation sign-off.

The workflow's value is not in its steps but in its discipline: every requalification decision is tied to a named validated channel, a cited piece of evidence, and a criticality level that determines who signs. That traceability is what allows a crash model to remain credible across a long design program without re-validating everything from scratch at every release.