Here are the general practices used in automotive crash models for defining contacts between occupants /vehicle, occupants /restraints, restraints/vehicle

All these will use SOFT=2 contact, SBOPT=3, DEPTH=35, PSTIFF=1. For contact between restraints and occupants/vehicle, restraints (bags and 2-D belts) are defined as slave and SST is set to ‘2.0mm’. In general, VDC (contact damping) is always used for sliding contacts in automotive simulations and a value between 30-50 is defined.

sp
1/29/2021

(email to LST Core from rg
Fri 1/29/2021 1:47 PM)

___________________________________________________

I strongly believe that Mortar contacts will see increased usage for localized/difficult contact situations in explicit.
It usually works the first time without having to tune any input variables. 

In a model from an Auto OEM user, the run time for his test case was around 15hrs on 96cores. 
After adding two local mortar contacts, the run time increased by by only 30 mins or so. 
In other words, the increase in cost was insignificant. 

(Paraphrased from sp)
Ticket#2019021210000063

______________________________________________________

Primary sources of information on contact in LS-DYNA:

http://www.dynasupport.com/tutorial/contact-modeling-in-ls-dyna
http://www.dynasupport.com/tutorial/ls-dyna-users-guide/contact-modeling-in-ls-dyna/
http//ftp.lstc.com/anonymous/outgoing/support/FAQ_docs/contact.pdf
http://ftp.lstc.com/anonymous/outgoing/support/FAQ_docs/contact_shorter.pdf
http://ftp.lstc.com/anonymous/outgoing/support/FAQ/contact*

A collection of LS-DYNA examples, including some contact examples, is available at
http://www.dynaexamples.com

______________________________________________

RE: Fabio's User Guide for Contact


Author: gunther
Date: 2019-12-04 16:56:02 -0800 (Wed, 04 Dec 2019)
New Revision: 11616
Added:
   trunk/LS-DYNA_Contact_Manual/
   trunk/LS-DYNA_Contact_Manual/Contact User's Guide.doc
   trunk/LS-DYNA_Contact_Manual/Contact_User's_Guide_002.doc
Log:
add contact manuals from Fabio (March 2006, 237 pages)


pdf is in outgoing/support/ContactUsersGuideINTERNALuseONLY.pdf

Ticket#2019120410000014

_________________________________________________________


The following are some very brief notes on LS-DYNA contact.  Not all the
contact types or contact options are discussed.  

Automatic vs. Non-automatic Contact:
Automatic contacts are recommended for most explicit simulations.  
Non-automatic contacts (in which 
contact orientation is important) are sometimes used for metal forming 
simulations where the 
geometries are very straightforward and contact surface orientation 
can be reliably established before the 
simulation is conducted.  
Non-automatic contacts are generally recommended for implicit simulations.
A major difference between AUTOMATIC and non-AUTOMATIC contacts for shells is that a 
non-AUTOMATIC contact will only be looking for penetration coming from one side of the shell.   
AUTOMATIC contacts have contact surfaces on both sides of the shell midplane.

More on Non-automatic Contact:
Shell offset consideration is optional in the non-AUTOMATIC contacts  and is controlled by:
SHLTHK in *CONTROL_CONTACT, and
THKOPT and SHLTHK on Card B of *CONTACT.
ORIEN on Card 1 of *CONTROL_CONTACT controls automatic re-orientation.


Contact types:
Type 13 contact (contact_automatic_single_surface) is a single surface contact 
(no master surface is defined) that always considers shell
thickness and has no orientation.  Thus it's necessary that shell surfaces
be modeled with at least a small gap between them.  To avoid initial penetrations,
the gap should be no less than the average thickness of the two shells potentially in
contact.  No gap is necessary between solid elements.
The contact searching algorithm for
type 13 contact is more complex than for type 3 (contact_surface_to_surface) or 
a3 (contact_automatic_surface_to_surface), i.e., type 13 can handle
situations such as shell edge to surface, and to some extent, beam to
shell surface.  As with any single surface contact type, resultant
forces are not directly retrievable in the RCFORC file; one must define a
contact_force_transducer_penalty in order to retrieve the contact forces.
The article "contact.13vs26" provides additional information on 
contact_automatic_single_surface and contrasts it to contact_automatic_general.

Type 3 contact (contact_surface_to_surface) contact is a 
surface-to-surface (two-way) contact where shell
thickness consideration can either be turned on or off either in *contact or
*control_contact (*contact has priority).
The orientation of the contact segments
is important with this contact type as the shell only looks in one direction for potential contact.

In a two-way contact such surface_to_surface, nodes on the slave side are first checked for
penetration thru the master surface and then master nodes are checked
for penetration thru the slave surface.  The exception is this approach is when
segment-based contact is invoked by setting SOFT=2.

Contact type a3 has no
orientation (a shell looks for potential contact from either side of the shell  
midplane) and always considers shell thickness, so in this regard,
it's quite similar to a type 13 contact.

Table 6.1 in the 950 Keyword User's Manual lists the maximum penetration d that defines
when a penetrating node is released from contact consideration.  This distance d is
different for a type 3 contact than for a type 13 contact.

Some notes on contact parameters:

SOFT
SOFT is the first parameter on Optional Card A of *contact....
The default value of SOFT is 0.  SOFT=1 is more or less the same as SOFT=0 
EXCEPT in the way the contact stiffness is determined.
SOFT=2 is a radical departure from SOFT=0, both in the way contact 
stiffness is determined but also in the manner that the search for penetration 
is conducted.  SOFT=2 invokes what is called  "segment-based contact". 
For notes regarding contact with SOFT= 1 and 2, see the articles "contact.soft1" 
and "contact.soft2", respectively.  

IGNORE
At any point during the simulation, if a node is suddenly found to be
below the surface (say, it was moving very fast and wasn't detected before
penetration), the old style (IGNORE=0) algorithm just moves the node to
the master surface without applying any forces (we term this "shooting node logic").  
If the shooting node logic is turned off (SNLOG=1),
then you get large forces suddenly appearing, and negative contact energy.
If IGNORE is set to 1 then the shooting node logic flag SNLOG has no affect.
Rather the amount of sudden penetration is noted and compensated for by
adjusting the contact thickness locally.
So at any time during the simulation, if a sudden penetration is
detected, the program doesn't apply any large forces nor are any nodes moved.
Contact forces, however, will resist FURTHER penetration. 


Question:
What value of SLSFAC should have in order to produce the physically adequate modelling of each problem ?

Answer:
SLSFAC is nondimensional scaling factor on contact stiffness that applies to all 
penalty-based contacts and is cummulative with any scaling factors applied in the 
*contact_... command.   If all of your contacts appear too soft, you might consider increasing 
slsfac and vice versa.   For explicit simulations, slsfac is rarely used since a better approach 
to scaling the stiffness is usually to set SOFT=1 (or 2).    
For implicit simulations, it's not unusual to have to increase (or decrease) SLSFAC 
by one or more orders of magnitude to get good contact behavior AND a converged solution.

The default value of SLSFAC of 0.1 is, based on experience, a good ballpark 
value for contact between stiff materials with about the 
same mesh refinement (in explicit simulations).   


Question:
For one model: a foam rectangular block hitting the aluminum plate (all parts
are modelled from solid elemts), I have different values of interface force depending
on SLSFC. With default value 0.1, it gives maximum of 5200 lbf, but with SLSFAC=0.5 it gives
maximum of 9100 lbf. The impact time interval is also reducing for the case of SLSFAC=0.5

Answer:
Such is the nature of penalty-based contacts in finite element analysis.   
Both the amplitude and duration of the force determine the 
impulse (area under the force vs. time curve).    
The impulse will be relatively insensitive for a fairly wide range of 
contact stiffnesses.   'Correctness' of contact behavior should be primarily 
judged on visual observation of how the 
parts interact with each other.  Some penetration is necessary to produce the 
contact force but the penetration should 
be small relative to element dimension soas to be nearly imperceptible while 
viewing an animation of the simulation.  
Contact energy (also called "sliding energy") offers additional 
insight into contact behavior. 


See also ftp://ftp.lstc.com/outgoing/support/FAQ/contact* 

jpd