3. Test matrix design

The knee mapping test matrix is the single most consequential planning artifact in a Euro NCAP knee mapping assessment. It determines which vehicles are physically tested, where the impactor strikes the front-end structure, and at what spatial density the resulting force–deflection corridors are sampled. A poorly designed matrix either wastes expensive sled and impactor time on redundant points or, worse, produces corridors that fail to capture the stiff local features — tow hooks, bumper beam ends, energy-absorber ribs — that dominate real pedestrian knee injury outcomes. This section explains how to build a defensible matrix from the vehicle selection stage through to the final grid resolution decision.

Vehicle model selection: representativeness over volume

Euro NCAP does not require every variant of a model family to be mapped. The assessment protocol expects the test laboratory to select vehicles that represent the front-end geometry and stiffness distribution of the model range sold in Europe. In practice, selection proceeds in three filters.

The first filter is structural family grouping. Vehicles sharing a platform and front-end module — same bumper beam part number, same bonnet leading-edge height class, same energy-absorber architecture — can normally be represented by a single mapping vehicle. The planner should obtain the OEM's front-end bill of materials and confirm that bumper beam, crash box, lower leg stiffener, and bonnet latch geometry are common across the variants under consideration.

The second filter is worst-case geometry selection. Within a family, the variant with the lowest bumper reference line and the most forward-protruding stiffener typically produces the most severe knee loading. Where the family spans a ride-height range (for example, standard versus sport suspension), the lower ride height is generally the conservative choice because it aligns the impactor with stiffer upper bumper structures.

The third filter is market coverage. If a model is sold with two distinct front fascias — for example, a standard and an off-road styling package with a different lower valance — these constitute separate mapping candidates even on a shared platform, because the fascia changes the effective contact surface the impactor sees.

Impact point definition and the vehicle coordinate reference

All impact points must be defined in a single, explicitly stated coordinate system before any testing begins. The convention used throughout this course is the vehicle coordinate system: X positive forward along the vehicle centreline, Y positive to the vehicle's left, Z positive upward, with the origin at the front axle centre at ground level. Every impact point is therefore recorded as an (X, Y, Z) triplet on the front-end surface, and the impactor axis is aligned with the local surface normal at that point.

Impact points are not chosen arbitrarily across the bumper face. The mapping grid is anchored to the bumper reference line (the horizontal line at the upper bumper surface used for legform testing) and to the bonnet leading edge. The vertical band of interest spans from the lower bumper valance up to the bonnet leading edge, because this is the region a pedestrian's knee and lower leg contact during a frontal impact. Points outside this band add little diagnostic value and inflate test count.

Horizontally, the grid must extend across the full width of the front-end structure that a pedestrian could plausibly contact, typically from Y = −700 mm to Y = +700 mm for a passenger car, but the planner should derive the actual limits from the vehicle's front-end width at the bumper reference line rather than applying a fixed number.

Grid resolution: balancing coverage against test count

Grid resolution is the spacing between adjacent impact points in the Y direction (lateral) and Z direction (vertical). Resolution is not a free parameter — it is constrained by the need to resolve local stiffness gradients. A bumper beam end, a tow-hook cutout, or a pedestrian-protection foam rib can change local stiffness by a factor of two or more over a lateral distance of 50–80 mm. A grid coarser than that will alias these features and produce a corridor that is artificially smooth.

A practical starting resolution for a passenger car front end is 100 mm lateral spacing and 75 mm vertical spacing, refined to 50 mm lateral in regions where the planner knows a stiff local feature exists. The refinement zones should be identified from the CAD model or a physical inspection of the bumper beam and energy absorber before testing, not discovered after the fact.

The total test count follows directly:

N_points = (Y_span / ΔY + 1) × (Z_span / ΔZ + 1)

Example: Y_span = 1400 mm, ΔY = 100 mm  → 15 columns
         Z_span = 300 mm,  ΔZ = 75 mm   →  5 rows
         N_points = 15 × 5 = 75 impact points

With a 50 mm refinement band of 200 mm width on each side:
         additional columns = 2 × (200/50 − 200/100) = 2 × 2 = 4
         refined N_points ≈ 19 × 5 = 95 impact points

Ninety-five impactor firings per vehicle is a realistic order of magnitude for a full mapping campaign and should be budgeted accordingly. If the programme cannot support that count, the correct response is to reduce the number of vehicles mapped — consolidating structurally equivalent variants — rather than to coarsen the grid below the feature-resolution limit.

Matrix structure and the coverage map

The complete test matrix is best represented as a coverage map: a plan view of the front end with the impact point lattice overlaid, colour-coded by whether each point is a baseline grid point, a refinement point, or a boundary point at the edge of the contactable region. The figure below shows this structure for a representative passenger car, including the refinement bands around the bumper beam ends and the tow-hook region.

The coverage map makes two things immediately visible that a text list cannot: the lateral extent of the contactable region, and the locations where the grid has been locally densified. It also exposes gaps — for example, if the refinement band around the tow-hook zone does not actually contain a refinement point, the map shows the omission before any hardware is fired.

Practical matrix decisions and common failure modes

Three matrix-design errors recur in assessment practice. The first is grid aliasing: a 100 mm grid that steps over a 60 mm-wide stiffener rib and reports a smooth corridor. The symptom is a corridor that is suspiciously flat compared with the physical structure; the fix is to inspect the CAD for ribs and add refinement points before testing. The second is coordinate drift: impact points defined on the CAD surface but applied to a physical vehicle whose ride height differs from the CAD nominal, shifting every Z coordinate by 10–20 mm. The fix is to measure the physical bumper reference line on the actual test vehicle and re-anchor the grid to it. The third is over-consolidation: treating two variants as structurally equivalent when their bumper beam part numbers differ, producing a corridor that represents neither variant well.

A defensible matrix is therefore one where every vehicle has a documented structural-equivalence justification, every impact point has an explicit (X, Y, Z) coordinate in the stated vehicle frame, and every refinement zone is traceable to a named local feature. When those three conditions hold, the resulting force–deflection corridors can be trusted as inputs to the assessment scoring, and the test count is justified rather than merely large.