RE: Issue of negative internal energy when using RECOV=10 jd Ticket#2019042910000144 https://groups.yahoo.com/neo/groups/LS-DYNA/conversations/topics/60734 James M. Kennedy Message 19 of 26 , May 3 Thanks for the information regarding mat_072r3. Please see the following presentations for "recove" discussions: Murray, Y.D., Abu-Odeh, A., and Bligh, R., "Evaluation of LS-DYNA Concrete Material Model 159", FHWA-HRT-05-063, Federal Highway Administration, Turner-Fairbank Highway Research Center, McLean, Virginia, May, 2007. http://www.fhwa.dot.gov/publications/research/infrastructure/structures/05063/05063.pdf 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 Engineering, University of Pittsburgh, Pittsburg, Pennsylvania, November, 2011. http://d-scholarship.pitt.edu/10611/1/Jaime_Maria_PhD_diss_ETD_Nov30_11.pdf Winkelbauer, B.J., Faller, R.K., Bielenberg, R.W., Rosenbaugh, S.K., Reid, J.D., and Schmidt, J.D., "Phase I Evaluation of Selected Concrete Material in LS-DYNA", MwRSF Research Report No. TRP-03-330-15, Midwest Roadside Safety Facility, University of Nebraska-Lincoln, Lincoln, Nebraska, April, 2016. https://mwrsf.unl.edu/researchhub/files/Report334/TRP-03-330-16.pdf glukrazor Message 21 of 26 , May 4 James, thank you for great references! I find no information about possible energy problem in these papers. As I understand, recov 1 and recov 11 makes the same thing - they disable brittle damage recovery capability. So, this is now very interesting scenario from concrete mechanics side. It seems to me, that now recov 10 feature is absolutely not applicable. It would be nice to talk with develpers about this material. avp2k Message 22 of 26 , May 4 This is a very interesting discussion. We have been using CSCM model for Road Safety related applications and usually ERODE was the only main parameter, which we needed to calibrate to achieve better agreement with the test results. James provided excellent references to recent publications, I have seen some of them earlier but decided to have a more detailed look after this issue with non-physical behavior. Influence of RECOVE has been studied in the Ph.D. thesis of Dr. Jaime for rock cutting application and value of 10.5 gave the best results. The main issue with CSCM model found in the recent MwRSF Research Report No. TRP-03-330-15 was the lack of repeatability of results using the same input parameters. As far as RECOVE parameter is concerned , they did not find significant change in the results, although they mainly considered values of 0.0, 0.5 and 1.0 of that input parameter. However, values of 10., 10.5 and 11 were also mentioned in the description of all the parameters studied. It is clear that CSCM model needs more validation studies for each new major application due to its complicated nature. James M. Kennedy Message 23 of 26 , May 5 Notes from below mention instabilities: Winkelbauer, B.J., Faller, R.K., Bielenberg, R.W., Rosenbaugh, S.K., Reid, J.D., and Schmidt, J.D., "Phase I Evaluation of Selected Concrete Material in LS-DYNA", MwRSF Research Report No. TRP-03-330-15, Midwest Roadside Safety Facility, University of Nebraska-Lincoln, Lincoln, Nebraska, April, 2016. https://mwrsf.unl.edu/researchhub/files/Report334/TRP-03-330-16.pdf 1. Finally, the recov parameter, which defines the recovery modulus in compression, was evaluated. This parameter was varied between 0 and 1, as well as 10 and 11. When set between 0 and 1, the recovery was based on the sign of the pressure invariant only. When set between 10 and 11, recovery was based on the sign of the pressure invariant and volumetric strain 2. Note that all cylinders showed different animated behavior, thus indicating randomness in the failure behavior and/or significant instabilities in the material model. glukrazor Message 25 of 26 , May 6 Erosion could helps to describe anything, but this feature does not allow you to predict results. I prefer not to use this feature as a real input. See the reference: https://cimec.org.ar/ojs/index.php/mc/article/download/3868/3786 ... EROSION CRITERIA FOR FRICTIONAL MATERIALS UNDER BLAST LOAD The conclusion from that reference says, "CONCLUSION A review of erosion algorithm frequently used in numerical simulations of blast and impact loading on concrete elements is presented. Erosion algorithms are numerical solutions to avoid great mesh deformation when Lagrange processors are used. Nevertheless, in the case of concrete they can be used to represent physical erosion, shear failure, cratering, spalling and fracture. The review presented shows a great dispersion of erosion criteria and erosion limits used by different authors for similar types of concrete. The differences found suggest that erosion limit requires further research. The review presented also shows that erosion criteria should be adequate to the type of “physical erosion” phenomena that is intended to be modeled. Although very simple, erosion criteria based on strain limits can be more easily related to physical phenomenon occurring in concrete under blast and impact loads. The review presented and the example developed show that this criterion is able to reproduce concrete failure under close blast loads. Numerical results are dependent on the erosion limit used. The use of erosion limits as high as possible is usually recommended in classical software used for the numerical simulation of blast or impact loads. The need of calibration with experimental results is also stated. What is not well established is if this erosion limit should be considered as a material property or not. The application example developed in this paper includes comparison with experimental results and proves that erosion limit is not independent of mesh size, thus it can not be considered as a material property. It is normally expected to obtain different results when the same problem is solved with different mesh sizes. Nevertheless the difference tends to disappear when the mesh is refined. If strain based erosion criteria is considered another type of mesh size dependency is introduced in numerical solution. Moreover, it should be observed that a very fine mesh must be used to obtain the shape of the damaged zone registered in tests under contact blast loads. This is not the case of more distant explosions characterized by flexure failure that can be modeled with coarser meshes. A simple correlation between mesh size and erosion limit is used in the paper to obtain similar numerical results. However, this dependency requires further research in order t establish the range of validity and if there are not other variables influencing this problem. __________________________________________________________________________________ I will be sending out code for mat159 from the R8 branch to Chris Galbraith of MFAC, intended for his customers BCHydro. Tod has received permission from John to release this to MFAC. I will send a zip file named mat159.zip, which contains a file mat159.F along with several include files and a README. The mat159.F file contains the following subroutines: qzmt159 Concrete159 sets159 inse159 f3dm159 Please allow this through your filters. Ushnish -- Ticket#2018120610000192 _________________________________________________________________ RE: Yvonne I reached out to Yvonne Murray, formerly of APTEK, [by forwarding] a message to her concerning the issue with MAT 159. She responded that she has been retired for about 3 years and does not have LS-DYNA or a copy of MAT_159 up on her computer. She also responded that the issue appears to be that the cap retracts below its initial position. Additionally, she has no distinct recollection of testing or having a problem with this issue one way or another. So, unfortunately, she cannot help. Jim Kennedy Ticket#2017060710000051 _____________________________________________________________ RE: rate effects With damage turned off in the CONCRETE-based input (B and D set to zero), I find less than a one percent rate effect on x-stress when comparing a strain rate of 1 vs. 20. I would call that difference insignificant. But when I increase the strain rate in the > second element to 200 (see 1vs200.k), the rate effect approaches 10% (final x-stress of -728 for epsdot=1 vs. -665 for epddot=200. I find it interesting and a little peculiar that the difference in stress (728-665=63) exceeds the input variable OVERC=21.64, the viscoplastic stress limit in > compression. > > As an experiment, I removed the transverse confinement (removed SPCs in y and z) to explore the effect of uniaxial compression. This shows the expected trend for rate effects, i.e., the fast element (epsdot=200) has a higher stress than the slow element (epsdot=1). To see this, run > 1vs200_uniaxial.k. Ticket#2014100810000065 ______________________________________________________________ enhancement request for d3hsp submitted 10/15/14 _______________________________________________________________________ As I read it, there is no user-control over the mesh regularization algorithm employed. For more on this, see Chapter 4 in the pdf at http://www.tfhrc.gov/safety/pubs/05063/index.htm and also see "Regulating Mesh Size Sensitivity" in the pdf available at http://www.fhwa.dot.gov/publications/research/infrastructure/pavements/05062/ref.cfm "Users Manual for LS-DYNA Concrete Material Model 159". _____________________________________________________________________ Element erosion in mat_159 can be based on (1) meeting only a damage criterion (set ERODE=1.0) -- OR -- (2) meeting both a damage criterion AND a max prin strain criterion (set ERODE > 1.0). Erosion does not occur if ERODE<1.0 unless the material is supplemented by *mat_add_erosion. As I understand it, the damage value considered (1) and (2) is max(brittle damage, ductile damage). In other words, both brittle damage and ductile damage are calculated separately and the worst of these two values is considered in evaluating element erosion (and element softening). Jim 2/15/11 _________________________________________________ I can only report I observe in the source code: When ITRETRC=1, x>hkmin and X0>hk0 as reflected in d3hsp: Miscellaeous Output Parameters: si1 Pressure apex of shear surface..........= -7.7451E+00 hkmin Minimum cap location ...................= 1.0000E-06 hk0 Initial cap location ...................= 2.1302E+01 hkcr Critical cap location ..................= 1.0000E+20 These 4 parameters are computed during initialisation: a) Find the intersection of the failure curve with the i1-axis: si1 is first computed by iteration using equation from Figure 19 with the inputs alpha, theta, lambda, beta. b) Find hk0 when x=0 hk0 is then computed by iteration using equation from Figure 19 with the inputs alpha, theta, lambda, beta and si1 computed in the previous step. c) hkmin=1.0E-06 d) if theta>0.0, hkcr =1.0E+20 If more clarifications are needed, somebody needs to contact Yvonne Murray, who developed this concrete model. Leslie 7/13/17 bug 13342 (submitted by Tobias) _________________________________________________ Details of *mat_159 beyond those in the User's Manual and Theory Manual are provided in the cited references, which you should be able to find online by searching on the FHWA report number: Refs: 1. Murray, Y.D., Users Manual for LS-DYNA Concrete Material Model 159, Report No. FHWA-HRT-05-062, Federal Highway Administration, (2007). 2. Murray, Y.D., A. Abu-Odeh, and R. Bligh, Evaluation of Concrete Material Model 159, Report No. FHWA-HRT-05-063, Federal Highway Administration, (2007). Yvonne Murray kindly sent me input decks that appear to correspond to the beam models mentioned in Chapter 6 of FHWA-HRT-05-063. I've placed these models in http://ftp.lstc.com/anonymous/outgoing/support/FAQ_kw/concrete/test_cases_from_mat159_report.tar . Be sure and read "readme1" therein. The examples in http://ftp.lstc.com/anonymous/outgoing/support/FAQ_kw/concrete/mat159_examples.tar appear to be from Chapter 8 of the report. _____________________________________ The bug in reporting of mat_159 damage in 2D simulations (bug # 2576) was fixed on 7/22/09. _________________________________________________________ the damage in fully-integrated elements was occurring only at the first integration point. The bugfix is available in the new release version R5.1. Reported by Ushnish 1/18/11. ____________________________________ Bug resulting in energy loss (#2269) fixed in 971 rev 52388. ____________________________________ The parameter IRATE in *mat_159 must be input as an integer. Note: LAMDA in the Users Manual is called "gamma" in the source code. Yahoo Yammerings in FEA Information newsletter (May 2008) states that fracture energy is independent of mesh size. This does not mean results are independent of mesh size but rather than the solution will converge to an answer as the mesh is refined. _____________________________ ***UPDATE*** The report "Evaluation of LS-DYNA Concrete Material Model 159" is now available online at http://www.tfhrc.gov/safety/pubs/05063/index.htm Sincerely, Akram Abu-Odeh Associate Research Scientist Safety and Structural System Division Texas Transportation Institute The Texas A&M University System 3135 TAMU College Station, TX 77843-3135 abu-odeh@tamu.edu Phone: (979) 862-3379 Fax: (979) 845-6107 "Users Manual for..." is available at http://www.fhwa.dot.gov/publications/research/infrastructure/pavements/05062/ref.cfm *** END UPDATE ******************************* Supplemental documentation of mat_159 (DRAFT copies of "Users Manual for..." and "Evaluation of..." mat_159) was received from Yvonne Murray, 12/2005. (Hardcopies on bookshelf).) Two input decks that accompany the "Evalulation" document are located in ~/test/mat159. Concerning the documentation, typically Martin Hargrave at the FHWA requests that draft reports be distributed as paper copies (and stamped DRAFT), until the FHWA is done with their review. So I am checking with Martin Hargrave to make sure it is OK for you to distribute paper copies. Once the review is complete, the reports will become available in both hardcopy and electronic form. For example, Bretts soil model 147 reports were completed by the FHWA last year and are available electronically: http://www.tfhrc.gov/safety/pubs/04095/index.htm http://www.tfhrc.gov/safety/pubs/04094/index.htm I expect the wood model reports to be available early 2006 in both hardcopy and electronic formats. We still have to go through another round of editing on the concrete model (mat_159, _cscm) reports, so they probably won't be available electronically until the end of 2006. Yvonne 12/15/05 --------------------------------------------------------------- 7/25/07 If IUNITS=4, the following error produced by 971 R2, rev. 7600.1116 ... *** Error Input error found *** Error reading material property cards ... is circumvented simply by running 971 R4 (still beta). I did not test 971 R3 (still beta, but next in line to be released). 971 R4 executables can be found in ftp://ftp.lstc.com/outgoing/ls971. If IUNITS=3, then FPC (if given) must be input in units of psi, .e.g., 4930 psi. This change alone resolves your second issue, even when running 971 R2,7600.1116. Furthermore, DAGG (if given) should be in units of inches and RO should be input in units of lbf-s^2/in^4.