4. Dummy positioning and initial conditions

In a MADYMO occupant-safety model, the dummy is not a passive passenger that simply sits where the CAD geometry places it. Its posture, the routing of every belt segment, and the initial velocity field assigned to the whole system determine the contact geometry, the belt-to-body load path, and the kinetic energy available at the first millisecond of the simulation. A model that is geometrically correct but kinematically inconsistent will produce belt slip, spurious penetrations, or an unrealistically early head strike. This chapter covers the practical techniques used in MADYMO pre-processing to establish a defensible initial state: posture definition, belt routing, and initial velocity assignment.

Coordinate Conventions and Reference Frames

Before any positioning work begins, fix the reference frame and keep it consistent throughout the model. MADYMO uses a right-handed Cartesian system; for a frontal occupant model the conventional vehicle frame is +X forward, +Y to the occupant's left, and +Z upward, with the origin typically at the vehicle datum. The dummy's own local frame is defined by its BODY system, and joint angles are expressed relative to the parent body. When you rotate the pelvis, the thorax and head follow through the kinematic chain, so a single pelvis rotation propagates to the whole upper body. Always state which frame an angle or velocity refers to; mixing vehicle-frame and dummy-local values is one of the most common sources of inconsistent initial conditions.

Posture Definition

Posture in MADYMO is controlled through joint degrees of freedom rather than by dragging geometry. Each joint (pelvis, lumbar spine, thoracic spine, neck, shoulder, elbow, hip, knee, ankle) exposes translational and rotational DOFs that can be locked, prescribed, or left free. The practical workflow is:

1. Place the dummy in the nominal seating position using the seat cushion and backrest as the primary contact references.
2. Adjust the pelvis rotation and the hip joint angles so the thighs rest on the cushion and the feet reach the toe pan without floating or penetrating.
3. Tune the lumbar and thoracic spine angles to set the torso recline, then the neck angle to level the head.
4. Set the upper and lower extremity angles to match the intended occupant posture (e.g., hands on the steering wheel rim, knees near the knee bolster).

For a 50th-percentile male Hybrid III in a driver position, a representative starting set is a pelvis angle of about 20–25° rearward from vertical, a torso recline of roughly 22–25°, and a head angle within a few degrees of level. These are example values for a mid-size sedan package; they must be verified against the actual seat geometry rather than copied blindly. After each adjustment, check for penetrations between the dummy surface and the seat, floor, and pedals. MADYMO's contact penetration output is the fastest way to detect a posture that is geometrically impossible.

Belt Routing

Belt routing is defined by a sequence of belt segments anchored at the retractor, D-ring, buckle, and anchor points, with the belt surface interacting with the dummy through contact. The key decisions are the number of segments, the friction coefficient, and the pre-tension. A typical three-point belt is modeled with a shoulder segment running from the D-ring across the chest to the buckle, and a lap segment from the buckle across the pelvis to the outboard anchor. Each segment is a BELT element with its own material and friction definition.

Routing technique matters more than most users expect. The belt must lie on the dummy surface, not float above it or cut through it. Start by defining the anchor and D-ring positions from the vehicle data, then let the belt settle onto the dummy by running a short initialization simulation with gravity and belt pre-tension active. Inspect the belt-to-body gap along the shoulder and lap paths; a gap larger than a few millimeters at the chest indicates the D-ring height or the shoulder angle is wrong. The friction coefficient between belt and dummy clothing is typically set between 0.3 and 0.5; a value that is too low allows the shoulder belt to slide off the chest during the crash, while an excessively high value suppresses realistic belt pay-out.

Initial Velocities

Initial velocity defines the kinetic state at t = 0. In a full frontal sled or vehicle model, the entire system is given a uniform initial velocity equal to the impact speed, for example 56 km/h (about 15.6 m/s) in the +X direction, while the barrier or deceleration pulse is applied to the vehicle structure. The dummy inherits this velocity through its parent system, so you do not assign it separately unless the dummy is modeled as an independent subsystem.

Two consistency checks are essential. First, the velocity direction must match the crash pulse direction; a sign error produces a dummy that moves backward relative to the vehicle. Second, angular velocities must be zero unless a pre-impact rotation is intentionally modeled. If the dummy is given an initial velocity different from the vehicle, the relative motion at t = 0 will generate artificial belt loading before the pulse begins.

INITIAL.VELOCITY
  SYSTEM  = dummy_system
  VX      = 15.6      ! m/s, +X forward
  VY      = 0.0
  VZ      = 0.0
  WX      = 0.0       ! rad/s, no pre-impact rotation
  WY      = 0.0
  WZ      = 0.0
END

After assigning velocities, run a zero-pulse settling check: with the crash pulse disabled, the dummy should remain essentially stationary relative to the seat for the first few milliseconds. Any drift indicates a residual force imbalance from belt pre-tension or contact over-penetration.

Verification and Common Failure Symptoms

The verification loop is short: settle, inspect contacts and belt gaps, confirm velocity consistency, then run a brief pulse. Only when the dummy remains stable through the first few milliseconds should the full crash simulation be launched.

The figure summarizes the three coupled inputs—posture, belt routing, and velocity—and the four checks that confirm a valid initial state. In practice, these are not independent: changing the torso recline alters the shoulder belt path, and changing the belt pre-tension alters the settling behavior. Iterate in the order posture, belt, velocity, then verify, and record the final joint angles, belt anchor coordinates, and velocity vector in the model documentation so the setup can be reproduced and audited.