Hello Nava –
I do not know of a publication that discusses the “evils” of using element erosion.
The problem with erosion is users tend to abuse its significant power. I sometimes use this handgun metaphor – when in the hands of a trained police officer it can save lives, but in the hands of an unknowing child it can inadvertently take lives.
So one cannot say *MAT_ADD_EROSION is “no good,” rather users who abuse *MAT_ADD_EROSION are computationally dangerous.
--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-dyna2/6b72f9de-db73-4b42-bf1d-486a905115ebn%40googlegroups.com.
Dear Nava,
Perhaps this might be helpful:
https://www.lstc.com/classes/fail_frac_damage
Element erosion: advantages & shortcomings
Sincerely,
James M. Kennedy
KBS2 Inc.
November 15, 2021
To view this discussion on the web visit https://groups.google.com/d/msgid/ls-dyna2/00a301d7da69%24ed838640%24c88a92c0%24%40schwer.net.
There are two aspects of erosion that mathematicians dislike in any numerical solution:
1\ Non uniqueness – the solution can be obtained in multiple ways. *MAT_ADD_EROSION offers many erosion criteria, and combined with an essentially infinite range of erosion criterion values, allows multiple ways to achieve the same eroded solution.
2\ Ad Hoc – solutions that work for one problem but not for a similar problem. If you have an eroded solution, just change the mesh discretization and see if you get the same result.
For some odd reason too many “computationally dangerous” users are not at all bothered by these aspects of erosion.
We are all “computationally dangerous” to some degree.
The least dangerous among us are those that know they are dangerous.
--len
To view this discussion on the web visit https://groups.google.com/d/msgid/ls-dyna2/313599897.1076846.1637055893110%40mail.yahoo.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/ls-dyna2/001601d7daf7%2449493230%24dbdb9690%24%40schwer.net.
The “algorithm” for *MAT_ADD_EROSION is simple, when any user specified criterion is met, or exceeded, the element is eroded. --len
From: ls-d...@googlegroups.com <ls-d...@googlegroups.com> On Behalf Of Kagan GENÇ
Sent: Wednesday, April 5, 2023 3:59 AM
To: LS-DYNA2 <ls-d...@googlegroups.com>
Subject: Re: [LS-DYNA2] *mat_add_erosion
Hi Len and Kennedy,
I would like to heat up this topic again since there are a lot of confusion. Some say there is no certain method to estimate these erosion criteria and we need to have a trial-and-error session to validate structural damage. I cannot process this statement in my brain, so I would like to ask you.
Let's think about this. There is an algorithm or a script works for erosion behind the interface. It is already determined I mean the way to delete element by this erosion algorithm. So, if we know how this algorithm deletes elements, I suppose we can estimate these values if we have material properties. At least, we should have known what the limits values(liek max and min) are for each erosion parameters as a result of this estimation. In my opinion, having the algorithm and material properties, we may obtain these erosion values.
So my question is that is there any documents present the scrip or the algorithm of mat_erosion?
Kind regards,
Kagan
17 Kasım 2021 Çarşamba tarihinde saat 20:14:58 UTC+9 itibarıyla Nava Farhadzadeh şunları yazdı:
Hi Oguz
I actually do agree with your statement. it will not affect the displacement history if the erosion value is not too low to cause premature failure.
Regards
Nava
On Wednesday, November 17, 2021 at 6:20:44 AM UTC kagan...@gmail.com wrote:
Hello everyone,
I want to share my own observations about erosion. I carried out several blast explosion simulations of a RC slab and found out that even if principal strain values in erosion were changed, the result displacement value did not change. I think erosion parameters should not affect results. Otherwise, this might be dangerous as you mentioned. Please take a look at the graph given below.
To view this discussion on the web visit https://groups.google.com/d/msgid/ls-dyna2/ed7abbd6-586a-4570-bc58-c926abc8aad6n%40googlegroups.com.
Hello Kagan –
Below is a posting I made to the old LS-DYNA forum in June of 2008 on the folly of using erosion. Perhaps this will answer your questions? --len
Many in this forum are familiar with my views on the use of erosion, i.e.
hate it!
Erosion is not a material characteristic, it is strictly a numerical artifact. There is no laboratory test to determine erosion anything.
The utility of erosion is to allow one continuum to pass through another when using a Lagrange solver. The availability of Eulerian and SPH solvers, for example, has obviated this utility, for the most part.
Still users continue with erosion. I can only assume out of a lack of motivation to learn new techniques.
But if you insist on using erosion here are three steps I consider
necessary:
1. Select an ad hoc erosion criteria; it really matters not what criteria since it is a numerical artifact. [BTW - you might as well change your eroded material to elastic and save plasticity computations, since your result will depend on a numerical artifact and not a material parameter.]
2. Perform an erosion criterion value convergence study, i.e. as the value of the selected ad hoc erosion criterion is changed how does the system response of interest change, and most importantly, does it converge to a non-trivial answer?
3. Given an ad hoc erosion criterion and converged erosion value, now perform a mesh convergence study. Again, as the mesh is refined how does the system response of interest change, and most importantly, does it converge to a non-trivial answer?
Steps 2 and 3 are inter-related and should be performed an iterative manner.
After all the above work, there is a fair chance you will discover there is no convergence. But even if you obtained a converged answer, there is no assurance it is correct, since you started with an ad hoc erosion criteria.
Some argue you can calibrate an erosion criterion to a given experiment.
That is true, but of what use, since you already know the experimental result; and again there is no unique way to perform the calibration, so what is the expected PREDICTIVE value of such a non-unique calibration.
If you like magic, you'll like erosion. But remember magic isn't mechanics.
--len
From: Kağan GENÇ <kagan...@gmail.com>
Sent: Monday, April 17, 2023 12:40 AM
To: L...@schwer.net
Subject: Re: [LS-DYNA2] *mat_add_erosion
Dear Leonard,
Thank you so much for your answer. The reason why I am thinking about this is if we know the algorithm it should not be that hard to assign the erosion parameters accordingly after investigating results. We can check max principal strain results and assign the parameters accordingly then. The only thing written everywhere about these erosion criterias is that there should be a trial and error session that needs to be done to assign parameters. So it does not make sense to me at all. We have inputs and outputs. Why does everybody carry out a trial and error session? or There is smt that I have not realized yet. That's all I wanted to debate about.
Kind regards,
Kagan
<l...@schwer.net> adresine sahip kullanıcı 12 Nis 2023 Çar, 03:58 tarihinde şunu yazdı:
To view this discussion on the web visit https://groups.google.com/d/msgid/ls-dyna2/002c01d9713a%24c38649e0%244a92dda0%24%40schwer.net.
By definition, when an element erodes it is removed from the computation. All that remains are the element’s nodes, i.e. representing portions of the element’s mass.
When adjacent elements erode, some of their common nodes may become unattached and act as rigid particles traveling with mass and velocity. These nodes may optionally remain active/inactive in the simulation as determined by the *CONTROL_CONTACT parameter ENMASS; see the User Manual Volume I. --len
Dear Suman,
A note taken from that may be of interest
Jaime, M.C., "Numerical Modeling of Rock Cutting and its Associated Fragmentation Process
Using the Finite Element Method", Ph.D. Thesis, Department of Civil and Environmental Eng-
ineering, University of Pittsburgh, Pittsburg, Pennsylvania, November, 2011.
http://d-scholarship.pitt.edu/10611/1/Jaime_Maria_PhD_diss_ETD_Nov30_11.pdf
5.3.1.1 Treatment of the Mass of Eroded Elements
The fact that elements are deleted from the finite element model is commonly thought as an
infringement of the basic modeling principle of conservation of mass. Fortunately, LS-DYNA
provides a mechanism to circumvent this violation. It is important to note that within the def-
inition of contact controls (see Section 5.2.2.3), the parameter ENMASS equals 1, which works
by retaining the mass of the eroded nodes in the calculation, and keeping them active in contact.
Consequently, the overall energy calculation of the system is not affected, as its mass is not re-
duced.
-------------------------------------
Some ENMASS notes:
https://ftp.lstc.com/anonymous/outgoing/support/FAQ_docs/contact_shorter.pdf
ENMASS - Flag for treatment of eroded nodes in contact.
An eroded node is defined as a node that is no longer attached to any element owing to element
deletion. To display eroded nodes as particles in LS-PrePost, toggle on “Show Deleted Nodes”.
ENMASS is not supported when SOFT=2 on Optional Card A of *CONTACT. ERODING_
NODES_TO_SURFACE is supported in both SMP and MPP. ERODING_SINGLE_SURFACE
and ERODING_SURFACE_TO_SURFACE are supported only in SMP.
EQ.0: Eroded nodes are not considered in the contact algorithm.
EQ.1: Eroded nodes of solid elements remain active in the contact algorithm.
EQ.2: Eroded nodes of solid and shell elements remain active in the contact algorithm.)
-------------------------------------
Sincerely,
James M. Kennedy
KBS2 Inc.
April 19, 2023
From: ls-d...@googlegroups.com [mailto:ls-d...@googlegroups.com] On Behalf Of SUMAN KUMAR
Sent: Wednesday, April 19, 2023 2:03 AM
To: L...@schwer.net
Hello everyone,
--
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/ed7abbd6-586a-4570-bc58-c926abc8aad6n%40googlegroups.com.
--
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-dyna2/002c01d9713a%24c38649e0%244a92dda0%24%40schwer.net.
--
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/CAFAuXadETGsJ_WKOi5bvFop1AS%2B%2BV9Y-xNLPNfG003pWemJnhg%40mail.gmail.com.