2. Target cascading principles

Target cascading is the discipline that turns a single, system-level crash performance requirement into a coherent set of engineering targets that subsystem and component teams can design against. In crashworthiness development, the top-level target is rarely a single number. It is a pulse — a time-history of deceleration measured at a defined location, usually the vehicle body at the seat track or the B-pillar base, during a defined impact event. The cascading problem is therefore: given a required full-vehicle pulse envelope and occupant response targets, what must the front-end structure, restraint system, seats, and body-in-white deliver, individually and together, so that the assembled vehicle meets the requirement?

The core principle is consistency under decomposition. A cascade is valid only if the union of the lower-level targets, when assembled and simulated, reproduces the upper-level target within an agreed tolerance. This is why target cascading is not merely "splitting up the work." It is a bidirectional process: top-down allocation of targets, bottom-up verification of feasibility, iterated until the two directions agree. A cascade that is never closed in the bottom-up direction is a wish list, not a target set.

The decomposition hierarchy

A practical crash-pulse cascade for a passenger car front impact typically has four levels:

  • Level 0 — Vehicle attribute target: occupant injury criteria (e.g., HIC15, chest deflection, femur load) plus structural intrusion limits, evaluated in the full-vehicle crash test at a specified speed and barrier type.
  • Level 1 — Full-vehicle pulse target: the deceleration–time corridor at the occupant reference point, plus the crash stroke (the usable longitudinal deformation distance) and the energy absorbed by the front structure.
  • Level 2 — Subsystem targets: the force–displacement characteristic of the front longitudinal rails, the crash box / bumper beam force level, the subframe contribution, and the restraint system firing schedule (pretensioner and airbag timing).
  • Level 3 — Component targets: rail wall thickness and material grade, trigger bead geometry, spot-weld and structural-adhesive layout, crash box crush initiator design, and mount stiffness.

The essential content of each level is a response characteristic, not a single scalar. A rail target expressed only as "absorb 60 kJ" is under-specified; the same energy absorbed at the wrong force level produces the wrong pulse shape and the wrong occupant loading. This is the most common failure in early cascades: energy and force level are treated as interchangeable when they are not.

Crash pulse target cascade from vehicle attribute to component targets Four-level hierarchy showing occupant injury targets at level 0, the full-vehicle deceleration pulse corridor and crash stroke at level 1, subsystem force-displacement and restraint timing targets at level 2, and component design parameters at level 3, with a bottom-up verification loop. Level 0 · Vehicle attribute target Occupant injury criteria: HIC15 ≤ 700, chest deflection ≤ 42 mm, femur load ≤ 7.6 kN Structural: toe-pan intrusion ≤ 120 mm, steering-wheel rearward displacement ≤ 90 mm allocate pulse + stroke Level 1 · Full-vehicle crash pulse target Deceleration corridor at seat-track reference point, 50 km/h full frontal rigid barrier Crash stroke ≈ 620 mm · total absorbed energy ≈ 118 kJ · peak deceleration ≤ 42 g a(t) t decompose force–displacement Level 2 · Subsystem targets Front longitudinal rails: mean crush force 165 kN, peak ≤ 210 kN, stroke 480 mm Crash box + bumper beam: 55 kN plateau, 120 mm stroke · subframe: 25 kN contribution Restraint: pretensioner at 12 ms, airbag full inflation by 32 ms, venting tuned to pulse F δ translate to design parameters Level 3 · Component targets Rail: DP780, t = 1.8 mm outer / 1.5 mm inner, trigger bead depth 6 mm, 45 mm pitch Joining: 32 structural-adhesive beads + 18 spot welds per rail · crash box initiator: 4 dents Bottom-up verification: assembled CAE / sled / full-vehicle test must reproduce Level 1 pulse within tolerance

Allocation is governed by physics, not by symmetry

The first quantitative step is an energy and momentum balance. For a full frontal impact at velocity v against a rigid barrier, the kinetic energy to be dissipated is E = ½mv², where m is the effective impacting mass. For a 1,700 kg vehicle at 50 km/h (13.89 m/s), E ≈ 164 kJ. Not all of this is absorbed by the front structure: some is retained as residual velocity of the occupant compartment, some is absorbed by the barrier and by friction, and some is dissipated in the rear structure after the front has bottomed out. A realistic front-structure share might be 70–75%, giving roughly 115–123 kJ — consistent with the 118 kJ example above.

That energy must be absorbed over the available stroke s. The mean force follows directly: Fmean = Estructure / s. With 118 kJ over 620 mm, Fmean ≈ 190 kN at the vehicle level. This is the number that must be partitioned among load paths. If the vehicle has two front rails, a subframe, and a bumper beam, the allocation might be 165 kN per rail pair (i.e., 82.5 kN per rail), 25 kN from the subframe, and the remainder from the bumper beam and lower load paths. The partition is not arbitrary: it must respect the stiffness and packaging of each path, and it must produce a pulse shape that matches the corridor, not just the correct area under the curve.

A useful sanity check at every level: the integral of force over displacement at the subsystem level, summed across all parallel load paths, must equal the vehicle-level absorbed energy within the cascade tolerance (typically ±5% on energy, ±10% on peak force). If it does not, the cascade is internally inconsistent regardless of how good each individual target looks.

Pulse shape as a cascaded requirement

The deceleration corridor is not a single curve but a band with an upper and lower bound at each time step. The upper bound protects the occupant from excessive peak loading; the lower bound ensures enough energy is absorbed early enough that the occupant does not run out of stroke. Cascading the corridor downward means translating it into a force–displacement target for the front structure. The relationship is approximately F(x) ≈ m·a(t) with x obtained by double integration of a(t). In practice this is done numerically: integrate the corridor bounds to obtain velocity and displacement histories, then map the force target onto the stroke.

Two consequences follow. First, the early portion of the pulse (roughly the first 20–30 ms) is dominated by the bumper beam and crash box, so their force plateau target is set by the corridor's early slope. Second, the mid-pulse plateau is dominated by the rails, so the rail mean force target is set by the corridor's central level. A cascade that assigns the same force target to both is almost always wrong: the crash box must be softer to initiate stable folding, and the rail must be stiffer to carry the main energy.

Coupling and the limits of independent allocation

Crash pulse cascading is not a clean tree. Several couplings break strict hierarchy:

  • Restraint–structure coupling: airbag and pretensioner timing are tuned to the pulse. If the rail target changes the pulse timing by 5 ms, the restraint firing schedule may need re-optimization. Restraint targets therefore cannot be finalized before the pulse is frozen, and the pulse cannot be frozen before restraint feasibility is checked.
  • Rail–subframe coupling: the subframe participates in the load path and can either assist or prematurely collapse the rail. Its target must be expressed as a compatible force–displacement, not an independent one.
  • Mass–pulse coupling: adding structure to meet a force target adds mass, which changes the kinetic energy and hence the required force. This is a fixed-point problem, usually resolved by two or three iterations.

Because of these couplings, the cascade is managed as an iterative loop with explicit interface variables. Each interface — pulse at the reference point, force–displacement at the rail rear mount, restraint timing — is owned by one team and frozen at a defined gate. Changes after freeze require a formal impact assessment against the upper-level target.

Closing the loop: verification and tolerance

The bottom-up leg is where most cascades fail in practice. Each subsystem team delivers a CAE model or test result for its component; these are assembled into a full-vehicle model; the resulting pulse is compared with the Level 1 corridor. Three outcomes are possible:

  1. Inside corridor, energy within tolerance: cascade is closed; targets are released for design.
  2. Pulse too high or too early: typically the rail force target is too high or the crash box is too stiff. The correction is applied at Level 2, not by relaxing Level 0.
  3. Pulse too low or stroke exhausted: the structure is bottoming out before absorbing the required energy. This usually indicates the stroke assumption was optimistic or a load path is not contributing as targeted.

A practical tolerance set for the cascade is: ±5% on absorbed energy, ±10% on peak deceleration, ±3 ms on pulse timing at key landmarks (e.g., time of peak, time to 50% of stroke). These are example values for a mid-size passenger car program and should be calibrated to the program's CAE correlation history. The important point is that tolerances are defined before decomposition, so that each level knows how much error it may contribute to the total budget.

Do not cascade a single scalar where a characteristic is required. "Rail must absorb 60 kJ" and "rail mean force 165 kN over 480 mm" are not equivalent targets: the first permits a force profile that violates the pulse corridor, the second constrains the shape. Always cascade the response characteristic — force–displacement, pulse corridor, timing schedule — and use scalars only as summary checks.

What a released cascade looks like

A completed cascade for one load path is a small, auditable package: the Level 1 corridor with its tolerance band; the allocated force–displacement target for the rail with mean force, peak force, and stroke; the corresponding component parameters (material grade, gauge, trigger geometry, joining pattern); the interface definition at the rail-to-body mount; and the verification result showing the assembled model reproducing the corridor. Each item is traceable upward, and each change is assessed against the level above. That traceability — not the elegance of the decomposition — is what makes the cascade usable by design teams who must commit tooling and validation budget against these numbers.

← Previous 📖 Contents Next →