1. MWD is dependent on the unit system of the model. You can consider this as an opposing pressure on the airbag. 2. There is no recommended MWD value. In general, its highly recommended not to use this function. It is definitely not a percentage value. Please use this following airbag (units:kg,mm,msec) to study MWD as you desire. I have run a few cases with MWD ranging from 0.2 to 200 in multiples of 10. MWD of 0.2 is very close to the case where no damping was used in terms of volume and mass flow rate out of the bag. The total energy in the system for MWD 0.2 has dropped considerably compared to no damping case. HD Ticket#2016122910000036 __________________________________________________________________ See http://ftp.lstc.com/anonymous/outgoing/support/FAQ_docs/airbag.ppt http://ftp.lstc.com/anonymous/outgoing/support/FAQ_docs/airbag052504_ian.ppt MWD [has units of 1/time and thus ] is unit dependent. for kg,mm,ms,kN - I have used 0.2-0.3 for Tonne,mm,sec,N - this should be 200-300. These values may be too high in some cases to where the unfolding of a bag is inhibited. A good value of damping will inhibit noisy flapping of the bag but still let the bag unfold in timely fashion. Too much damping would put very large external forces on the structure. Dilip has seen people using very high damping in tires for 5mph bumper tests and wondering why the vehicle slows down without a barrier. __________________________________________________________________ The reference geometry should represent the physical shape of a completely unstressed airbag. Think of it as a deployed (unfolded) but unstretched configuration. Use of a reference geometry (supposedly) prevents nonphysical stresses from developing as a result of having nonphysically distorted elements in the initial, folded geometry. __________________________________________________________________ Before trying to run your data, I've looked at it and can make some suggestions to hopefully improve it. The first point has to do with the fabric material liner. In general, people use an airbag liner that is about 10% of the fabric thickness (LRATIO=0.1) and a liner elastic modulus equal to the modulus of the material. In addition, if the airbag has a reference geometry, it usually is helpful to turn off compression in the liner until the reference geometry is reached by setting LNRC=1. In other words, the liner will not be active initially but will become active after elements are stretched to their reference dimensions. In the data you sent, the liner is very thin and has a small stiffness so is likely nearly ineffective except that it generates compressive stress initially causing the bag to open prematurely. The rest of this note has to do with airbag contact. It my experience, it does not help to scale up the airbag contact thickness, and it can cause stability problems because too many segment come into contact. Therefore, I would suggest that you leave SFST, and SFMT at their default values. Also, I suggest that you not set EDGE=5 on optional card A. For airbags, the recommended value of EDGE is 3 for version 970 and 0 for version 960. Setting EDGE=5 in 970 causes the 'sliding' option to be active. This option helps mainly with the sliding of solid parts and is not helpful for airbags. Using either version 960 or 970, airbag unfolding is usually more accurate if you set DEPTH=3 on optional card A. However, its not clear if more accurate contact is most effective at preventing tangling of the bag. In the data you sent, the DEPTH parameter is 0 which causes the default value of 2 to be used. The default value can be less accurate and cause more jagged looking airbag segments, but may more effectively prevent tangling of the bag. Finally, a small solution time step is important for airbag contact. I see that you set the time step to 2 microseconds which may be small enough. If airbag layers penetrate too much, you may need to further reduce to 1 microsecond or even 1/2 microsecond. A smaller time step causes the contact stiffness to be greater and should result in less penetration. The square of the time step appears in the denominator so reducing the time step by 2 causes the stiffness to increase by 4. LPB > However we are facing a new problem, the airbag doesn't look like it is > deploying completely. If you can help us on this, it would be greatly > appreciated. > revised 2/2004 jpd ref geometry (not in o2a version of this file)