Nonreflecting Boundary conditions in ALE
Arumugam ce19d044
l...@schwer.net
The present implementation of nonreflecting boundary conditions is intended for elastic waves, and thus not appropriate for air blast.
The only known solution to eliminate air blast boundary reflections is to increase the distance from the boundary to the region of interest.
--len
--
You received this message because you are subscribed to the Google Groups "LS-DYNA2" group.
To unsubscribe from this group and stop receiving emails from it, send an email to ls-dyna2+u...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/ls-dyna2/CAG%2BwLY9sFuGmKOEnowoT_DH4BCMU4NZVzXqzLOBkXQwJi1Y0Yw%40mail.gmail.com.
James M. Kennedy
Dear Arumugam,
A note I received from LSTC.
FYI: it will be possible to provide 2D *SET_SEGMENT for *BOUNDARY_NON_REFLECTING_2D
in later revisions of LS-DYNA (from R14). I did the test for your model and it ran to
completion (i.e. the *SET_NODE for *BOUNDARY_NON_REFLECTING_2D could be responsible
for the time step drop that I saw lately)
-----------------------------------------
A note I found posted by Len Schwer which questions the validity of simulations
when mixing blast with non-reflecting boundaries:
1\Non-reflecting boundaries will not maintain the preload stress as they only work
against velocities and not displacements. Also, non-reflecting boundaries are not
effective for blast simulations as they are intended to prevent the reflection of
“elastic” waves, which are quite small compared to blast waves.
More notes by LES:
1\ I am not sure why you want to use the MM-ALE solver for a wave propagation problem,
but be advised I think the *BOUNDARY_NON_REFLECTING only works with 2D & 3D Lagrange
solids.
2\ The *BOUNDARY_NON_REFLECTING feature is only intended to work with elastic waves.
3\ You could surround your MM-ALE mesh with a few layers of Lagrange solids and then
use *BOUNDARY_NON_REFLECTING on the Lagrange layers.
-----------------------------------------
Sincerely,
James M. Kennedy
KBS2 Inc.
May 16, 2022
From: ls-d...@googlegroups.com [mailto:ls-d...@googlegroups.com] On Behalf Of Arumugam ce19d044
Sent: Sunday, May 15, 2022 11:46 PM
To: ls-d...@googlegroups.com
Subject: [LS-DYNA2] Nonreflecting Boundary conditions in ALE
Dear All
--
Kağan GENÇ
I have been modeling the simulation of the RC slab experiment under blast load. For that reason, I firstly verified the blast pressure load measured from a particular point in the model as done in the experiment. The results agree well. Also, for another study, I tried to use non-reflective BC because of the need for fine mesh usage and the size of the model. Additionally, I have read some articles in the literature, which non-reflective BC are used for blast loading in. To verify this knowledge, I created a bunch of blast pressure test simulations using the MM-ALE method and I managed to get rid of the reflections of pressure waves from boundaries by using this keyword. What I have observed is that even if there is a little energy loss and the first peak of pressure values is postponed a little bit which can be seen from the chart below, the difference between the values for my case does not exceed 5%, which seems acceptable to me. So I have been thinking of using this keyword(non-reflective BC) in my new bigger model but I have seen you and Mr. Kennedy's comment on the issue, which emphasizes the inability of this keyword for blast loading due the limitations for elastic waves. Based on my study, the results do not seem bad in my opinion but I am curious about your comments on the chart below.
To view this discussion on the web visit https://groups.google.com/d/msgid/ls-dyna2/000201d86912%243411f6f0%249c35e4d0%24%40schwer.net.
l...@schwer.net
Hello Kagan –
I never argue with success, but I am not sure I fully understand your pressure histories.
I assume this is the pressure history on, or near, your structure. In one case there is no nonreflecting boundary condition and in the other case you have activated nonreflecting boundary conditions.
If the above statement is correct, there seems to be little if any reflection from the nonreflecting boundary, but I do not understand how changing a distance boundary condition can delay the time of arrival of the blast or change the maximum pressure?
--len
BTW – I doubt the units on your chart are MPa x 10^6?
From: Kağan GENÇ <kagan...@gmail.com>
Sent: Tuesday, May 17, 2022 1:08 AM
To: L...@schwer.net
Cc: LS-DYNA2 <ls-d...@googlegroups.com>
Subject: Re: [LS-DYNA2] Nonreflecting Boundary conditions in ALE
Dear Leonard,
I have been modeling the simulation of the RC slab experiment under blast load. For that reason, I firstly verified the blast pressure load measured from a particular point in the model as done in the experiment. The results agree well. Also, for another study, I tried to use non-reflective BC because of the need for fine mesh usage and the size of the model. Additionally, I have read some articles in the literature, which non-reflective BC are used for blast loading in. To verify this knowledge, I created a bunch of blast pressure test simulations using the MM-ALE method and I managed to get rid of the reflections of pressure waves from boundaries by using this keyword. What I have observed is that even if there is a little energy loss and the first peak of pressure values is postponed a little bit which can be seen from the chart below, the difference between the values for my case does not exceed 5%, which seems acceptable to me. So I have been thinking of using this keyword(non-reflective BC) in my new bigger model but I have seen you and Mr. Kennedy's comment on the issue, which emphasizes the inability of this keyword for blast loading due the limitations for elastic waves. Based on my study, the results do not seem bad in my opinion but I am curious about your comments on the chart below.
James M. Kennedy
Dear Arumugam,
A note I received from LSTC.
FYI: it will be possible to provide 2D *SET_SEGMENT for *BOUNDARY_NON_REFLECTING_2D
in later revisions of LS-DYNA (from R14). I did the test for your model and it ran to
completion (i.e. the *SET_NODE for *BOUNDARY_NON_REFLECTING_2D could be responsible
for the time step drop that I saw lately)
-----------------------------------------
A note I found posted by Len Schwer which questions the validity of simulations
when mixing blast with non-reflecting boundaries:
1\Non-reflecting boundaries will not maintain the preload stress as they only work
against velocities and not displacements. Also, non-reflecting boundaries are not
effective for blast simulations as they are intended to prevent the reflection of
“elastic” waves, which are quite small compared to blast waves.
More notes by LES:
1\ I am not sure why you want to use the MM-ALE solver for a wave propagation problem,
but be advised I think the *BOUNDARY_NON_REFLECTING only works with 2D & 3D Lagrange
solids.
2\ The *BOUNDARY_NON_REFLECTING feature is only intended to work with elastic waves.
3\ You could surround your MM-ALE mesh with a few layers of Lagrange solids and then
use *BOUNDARY_NON_REFLECTING on the Lagrange layers.
-----------------------------------------
Sincerely,
James M. Kennedy
KBS2 Inc.
May 17, 2022
From: ls-d...@googlegroups.com [mailto:ls-d...@googlegroups.com] On Behalf Of Arumugam ce19d044
Sent: Sunday, May 15, 2022 11:46 PM
To: ls-d...@googlegroups.com
Subject: [LS-DYNA2] Nonreflecting Boundary conditions in ALE
Dear All
--
Kağan GENÇ
I have just written wrong as MPa, it is Pa. I have just modeled this simulation to verify the blast pressure value so there is no structure that the blast load can collide at first as can be seen on the pic below. In addition, the three surfaces without reflected BC are defined as segments to use non-reflective BC keyword.
Arumugam ce19d044
To view this discussion on the web visit https://groups.google.com/d/msgid/ls-dyna2/000201d86912%243411f6f0%249c35e4d0%24%40schwer.net.
l...@schwer.net
The LS-DYNA implementation of nonreflecting boundary conditions is a linear viscous damper.
The idea is to apply an external stress (pressure) on the outer boundary that is equal to the stress in the wave arriving at the boundary, thus making the finite boundary appear infinite.
In an ELASTIC medium, the relation between stress and velocity in a stress wave is: Sigma = rho * wave speed * velocity
Where elastic wave speed is sqrt(modulus/rho)
In a blast wave, the density (rho) is not constant and the wave speed exceeds the elastic wave speed and is dependent on the pressure (stress) in the blast wave.
So how would you propose to know a priori the relation between stress and velocity for a blast wave?
To view this discussion on the web visit https://groups.google.com/d/msgid/ls-dyna2/CAG%2BwLY9j1HL3oWXhKcFZJo2Vz7QbvMp9j_SYzKXQ8upcy5%3D0GA%40mail.gmail.com.