PRIMER 22.1

A Mini-Tutorial on PRIMER Seatbelt Fitting

A Mini-Tutorial on PRIMER Seatbelt Fitting

The seatbelt fitting routines are designed to fit one or more belts to a structure, using shell and 1d and/or 2d seatbelt elements. It is also possible to define the belt-related items such as retractors , sliprings, sensors , etc, and to specify database cross-sections down the belt.

While intended primarily for Occupant analyses, the "structure" used for belt fitting does not have to be a dummy: it can be any combination of solid, shell and thick shell elements, and belt fitting may be used in any context where a line of elements needs to be fitted around an arbitrary shape. In order to create and fit a seatbelt you need to go through the following process:

(T1) Define the components to which the belt will be fitted.

The first step is to define those elements of dummy and nearby structure against which the belt will be fitted. Since fitting (form-finding) is an iterative process that involves contact calculations it is worthwhile trying to minimise the number of elements to be considered, since this will speed up the process.

The image below shows a very simple sled model in which those elements which need to be considered for belt fitting, ie those which the belt will contact, have been coloured red (dummy) and green (nearby structure). Dummy elements that need not be considered are blue, and structure that need not be considered is grey.

During the actual analysis the belt may come into contact with airbag, B-Post, steering wheel and other components, and the contact surfaces used for the Ansys LS-DYNA analysis itself should consider this. However during belt fitting only contact with the red and green elements needs to be considered, and the process will be faster if only these are used.

( It might be reasonable also to exclude the dummy's upper leg elements in this case, however experience shows that when seat positions are moved back and forth it is possible for the upper legs of dummies to contact the lap section of the belt during fitting, even if the final shape does not pass over them, so it is usually sensible to include the upper legs in the contact as shown here.)

For a more detailed description see Define: Defining "Structure" for seatbelt fitting

(T2) Define the basic path of the belt.

The image on the left below shows the basic path, made up of 8 points located at nodes on the structure or dummy.

  • Points 1 and 2 are on the B post
  • Points 3 and 4 are on the dummy's chest
  • Point 5 is at the pelvis buckle position
  • Points 6 and 7 are on the dummy's pelvis
  • Point 8 is the final attachment point on the floor-pan.

The belt fitter creates a curved path through these points, with breaks where the angle is acute, the black line in the left hand image. This is an example of a "simple" path, typical of front and side impact adult dummies.

Then it projects this path forwards from the dummy, calculates the plane of the belt elements, and creates an initial meshing path as shown in the right hand image. At this stage the meshing path is some way forward from the dummy, and not in its final shape, for example the lap section is far too high above the pelvis. This does not matter at this stage, the important thing is that the meshing path should be topologically "outside" the dummy and not have any initial penetrations into it.

For a more detailed description see Creating a belt "path"

Basic path, here using 8 points

Meshing path derived from this.

(T3) Add detail to and adjust the meshing path.

The left hand image shows typical additions to the basic belt path which add detail.

  • A retractor has been defined at point 1 on the B-Post
  • Sliprings have been defined at points 2 (top of B-Post) and 5 (pelvis buckle)
  • Database cross-sections to obtain cut forces during the analysis have been added at points 2, 4 and 5

The right hand image shows how the angle and position of both sliprings has been adjusted to give an accurate match with the geometry at those points. (Points can be moved and the twist of the belt can be adjusted to obtain a better initial shape.)

For more information see Adding detail to the belt path .

The fully configured meshing path before fitting

Note! It is also possible to define "fully meshed" sliprings where 2d belt elements or shells are meshed continuously around a radiused slipring, see below.

(T4) "Fit" and then mesh the final shape.

The left hand image shows the final shape of the meshing path after "fitting". This is a form-finding process in which the belt is pulled inwards onto the dummy until it makes contact, and it can also slide across the dummy in order to try to find the shortest path between ends. In this example the shape of the chest section has not changed very much, but observe how the pelvis section has pulled down onto the lower torso to achieve a realistic shape.

The right hand image shows the final part of the process: meshing the fitted belt. In this example the mesh consists of 1d belt elements between retractor and slipring, and also a short stretch of 1d belt elements either side of both sliprings. The chest and pelvis sections are meshed with fabric shell elements.

For more information see FIT: Commencing the form-finding operation

The fitted meshing path after form-finding

The final meshed result, here 1d belt elements and shells

PRIMER can mesh with a mixture of 1d belt elements, 2d belt elements and traditional shells, and it is easy to remesh a fitted path at will with a different combination of element types. For example the left hand image below shows the same fitted meshing path remeshed with 2d belt elements for the whole belt. The mesh is now continuous through the sliprings, with all the various node and element sets required by 2d sliprings and retractors created as required.

It is also possible to replace discrete slipring elements (*ELEMENT_SEATBELT_SLIPRING) with explicitly meshed shells or 2d seatbelt elements wrapped around the specified radius. The right hand image below shows such a mesh at the shoulder location in the model above. Additionally, a new option was added in version 20 that automatically pulls an explicitly meshed shoulder slipring into its correct position during the fitting process (see Explicit Slipring Panel for more details). 

Remeshing the result above entirely in 2d belt elements

Detail of an explicitly meshed (radiused) D-Ring

PRIMER also allows you to define the attributes of retractors, sliprings, sensors, pre-tensioners, etc and these can be "remembered" when the belt is remeshed so that they don't have to be repeatedly redefined.

(T5) Refitting belts automatically.

The first time you fit a belt to a dummy you will need to define the path and mesh the result manually.

However it is often the case that a dummy needs to be moved, typically when adjusting the seat position, and then remeshed in its new location. Because the basic path points in section (2) above are located at nodes on the dummy and structure PRIMER "knows" when these nodes move, so it can update the basic path shape using their new coordinates. From this revised basic path it can generate a new meshing path and refit this, then using the existing meshing information it can remesh the revised belt path.

The Auto-Refit function performs the following tasks:

  • Adjusts the fitting path according to the movement of nodes at which path points are defined.
  • Refits the belt to the new dummy position using this revised path.
  • Deletes the old belt and remeshes it reusing the original properties, part ids, sliprings, retractors, cross-sections etc

So long as the belt path is reasonably "simple", using at least some nodes on the dummy to locate path points, and the dummy has not moved such a long way that the original path can no longer fit successfully this process can be performed automatically in a single operation. More complex paths, and in particular those where most or all path points are no longer associated with nodes on the dummy, cannot be refitted this way since the process uses the motion of nodes on the dummy to move initial projected path to its new position.

For more information see Auto-Refit : Refitting a belt automatically when the dummy moves

(T6) Replacing one dummy with another.

Another common operation is to replace dummy A with dummy B, perhaps replacing a male dummy with a female one. The two dummies will have roughly the same shape and organisation, but the details of their meshing will be different and it is unlikely that a node number used to locate a belt path point on dummy A will be in the same place on dummy B.

PRIMER cannot perform this process automatically, but it can make the manual task easier

The first step is to define the part and / or element sets that make up the new dummy structure against which the belt will be fitted.

PRIMER will check these (left hand image above) then will attempt to identify nodes on the new dummy that are close to path points of the existing basic path (right hand image). Once a sufficient number of path points have been found Update dummy and belt path will replace the old dummy definition with the new one.

For more information see section New Dummy : Replacing one dummy with another

(T7) Fitting belts to more complex shapes.

So far this tutorial has explained the basic process of setting up, fitting and meshing an adult dummy model since this is the simplest case.

However PRIMER can also fit belts to more complex models such as a child dummy in a seat with guides and "wings" through which the belt must pass. The procedure is similar to the above, but the definition of the basic belt path is more complex. The basic path editor has various preset modes, one of which is "child dummy in seat", which helps with this process.

This example shows two views of a typical child dummy in a seat, with the belt guide detail at the shoulder enlarged.

Child dummies such as the example above do not conform to the simple "rubber band around an egg" shape that can be used for adult dummies. The belt path can undergo reverse curvature, it may have to thread through holes and guides, and in the region of the child seat "wings" it may have to navigate very tight geometry. There are several consequences arising from this:

  1. Defining a simple path at nodes, projecting it forwards and then pulling it back is not going to work.

    This means that while path points may start of being located at nodes it is likely that they will need to be moved away manually into "thin air", breaking the relationship with the original node. This does not affect manual fitting, but it does mean that "auto refit" after adjusting the dummy's position will not work if the dummy has moved any significant distance.


  2. Trying to calculate belt path twist according to nearby structure will tend to give the wrong answer near complex geometry

    The adult belt fitter tries to align the initial belt path "twist" with the nearby structure, but for child dummies this method can be replaced with a simpler approach based purely on the basic path curvature. This does mean that more manual intervention will be required to control the belt twist in critical regions.

Fitting a belt to a child dummy in a seat, and to other complex geometries, is more labour-intensive than fitting to the simple adult case, but PRIMER has various tools which help with the process. The Fitting options panel contains many feature that will help with this process.

Advanced belt path creation

Version 18 introduces an "advanced" belt path creation mode which improves dramatically the process of fitting a belt path to complex shapes. It includes

  • Far more control over belt curvature and direction, including the ability to "break" the natural curvature of the path so that it can move in any direction.
  • Much improved graphics and interactive path manipulation with the mouse.
  • The ability to vary belt element length and friction in localised regions of the belt path.
  • The ability to "curl" the shape of the path to help it to fit into awkward geometries.

This makes it possible to fit a belt to arbitrary shapes, including fully meshed buckles and D-rings, guides and generally tight and "difficult" geometry.

The Advanced fitting section describes this in more detail.

(T8) Saving results in the keyword output file.

$
*BELT_START
1Seatbelt definition 1
0 700 0 0 1113
4 40.0 15.0 3.0 200.0 150.0 0.0 0.0
25 1.0E-5 0 1.0 5.0 2.0E-3 25.0 30.0
0.0 90.0
2 2 0 0 4 0
1 1.0 500 5.0 15.0 20.0
$
*BELT_MESH
1053 0 709 2349.002 1 3
0 0 1 1 1 1 1 5
1045 1051
3000 0.0

PRIMER has its own set of keywords, prefaced *BELT, which describe each seatbelt definition in a model. Since these are not Ansys LS-DYNA keywords they are written after *END in the relevant file.

They save all the information necessary to define, modify, refit and remesh a belt path. Details of the card formats are given in Appendix E .

(T9) Using *AIRBAG_REFERENCE_GEOMETRY for "all shell element" belts

From Version 18 it is practical to create fully meshed belts using shells (as opposed to 2d seatbelt elements) which makes it possible to define the belt wholly from *MAT_FABRIC, allowing genuine fabric material properties to be used.

PRIMER attempts to keep belt meshes reasonably regular and orthogonal, but this becomes difficult when explicit shell meshes negotiate complicated geometries where in real life some "scrunching up" of the belt material would occur. As a result the as-fitted belt may not have a totally regular shape, that is if you removed it and laid it out straight the shape might be a bit "wobbly" with small variations in width and minor kinks, which is not 100% realistic.

To get round this Version 18 is capable of generating *AIRBAG_REFERENCE_GEOMETRY cards for the belt. This keyword was introduced to Ansys LS-DYNA in order to deal with a similar "initial distorted shape" problem when airbags are folded, and it means that the final deployed shape of the airbag will not be affected by these distortions. Seatbelts face exactly the same problem and so long as *MAT_FABRIC is used with conventional shell elements (*ELEMENT_SHELL) then the final shape of the belt will be correct and spurious stresses will not arise from initial distortions. See Create Reference Geometry for more details.