RE: SPG in modeling concrete or rock As an alternative to using Lagrangian solids or ALE or SPH, SPG has application to modeling of concrete or rock. There is a document for impact penetration on concrete targets using SPG at http://ftp.lstc.com/anonymous/outgoing/ycwu/SPGdocs/SPG4Conc.pptx The keywords for the examples in the document can also be downloaded at http://ftp.lstc.com/anonymous/outgoing/ycwu/Keywords/impsqslab.tar.gz http://ftp.lstc.com/anonymous/outgoing/ycwu/Keywords/imp400mmcyl.tar.gz http://ftp.lstc.com/anonymous/outgoing/ycwu/Keywords/imp800mmcyl.tar.gz ___________________________________________________________________ For examples of material input for concrete, see the bottom of this file. For a list of some references pertaining to concrete materials, see http://ftp.lstc.com/anonymous/outgoing/support/FAQ/concrete_references To search for papers addressing LS-DYNA applications, go to www.dynalook.com. Here are a couple: Schwer, L.E., and Malvar, L.J., "Simplified Concrete Modeling with *MAT_CONCRETE_DAMAGE_REL3," JRI LS-DYNA User Week, Nagoya, Japan, November, 2005. http://www.geomaterialmodeling.com/pdf/186concrete.pdf http://www.geomaterialmodeling.com/pdf/88briefly.pdf Schwer, L.E., "Simple Input Concrete Constitutive Models: An Illustration of Brick Walls & Concrete Cylinder Perforation", 10th International LS-DYNA User's Conference, Dearborn, Michigan, June, 2008. http://www.dynalook.com/international-conf-2008/PenetrationBlast-4.pdf/ Dr. Len Schwer offers a short course in Geomaterial Modeling geared toward LS-DYNA users. See the class schedule under www.lstc.com Concrete falls under the heading of geomaterials and thus discussion of concrete modeling is included in the course. If you're interested in a copy of the course notes and/or arranging for an on-site course, contact Dr. Schwer via email to Len@Schwer.net. For more information, go to www.geomaterialmodeling.com. When little is known about the concrete material properties, consider mat_072R3 (preferred over regular mat_072), mat_016, and mat_159 as these have options to generate material constants given the unconfined compressive strength as input. Have a look at this summary of these three concrete models: http://www.geomaterialmodeling.com/pdf/88briefly.pdf http://www.geomaterialmodeling.com/pdf/186concrete.pdf <-- PowerPoint As of R6.0.0, mat_272 also has an automatic input generation capability. These generated parameters are written to d3hsp in R6.1.1. More from Len Schwer regarding the "simple input" concrete models (6/10/2010) ... Cylindrical sample unconfined compressive strength values should be used for the MAT159. You raise an important question for what I call 'simple input' concrete models, e.g. MAT016, MAT072R3, MAT084/85, & MAT159. Each of these models should explicitly state if the compressive strength value is for cylindrical or cubical samples. I believe only MAT084/85, on my short list, uses cubical samples. (more details in "mat84_winfrith" notes) BUT... on 9/24/2010, Martin Bleachley of Serco, in regard to MAT84/85, said "As the cylinder strength of concrete < cube strength, it is usual to use the former. Although I cannot say for sure, it is likely to be the cylinder strength as we have documentation validating the model (Load v strain) against a concrete cylinder test." Rate effects in concrete: Observed behavior at high velocities can lead one to overestimate the strain rate effect in concrete. Such behavior is likely more attributable to inertial confinement rather than to material rate effects. _____________________________________________________________________ RE: Further discussion of material models for concrete in LS_DYNA Shells and beams: A resultant beam approach to modeling of reinforced concrete beams and columns may be employed in some cases using *mat_139 or *mat_191. For an integrated concrete beam formulation, see mat_195. *Mat_172/*mat_concrete_ec2 is a plain or reinforced concrete model for shells and H-L beam elements which includes thermal effects. See *Mat_203/*mat_hysteretic_reinforcement as an option for defining reinforcement when using *mat_172 for the concrete. See http://ftp.lstc.com/anonymous/outgoing/support/FAQ_kw/concrete/MAT_172_notes_03apr2018.pdf *Mat_174/*mat_rc_beam is a plain or reinforced concrete model for H-L beams that includes cyclic behavior which is important for seismic analysis. *Mat_194 (*mat_rc_shear_wall) can be used with shells to model "squat shear walls" or concrete slabs. Solids: Material models 5, 14, 16, 25, 72, 72r3, 84/85, 96, 111, 145, 159, 272 can be used to model concrete with solid elements. Some sample input for concrete is given for mat_145 in the User's Manual. A third cap model, mat_159 (*mat_cscm(_concrete)) has some built-in parameters for concrete (see also http://ftp.lstc.com/anonymous/outgoing/support/FAQ_kw/concrete/mat159 ). Mat 16 (Mode 2) is convenient if you know the unconfined compressive strength of your concrete but don't have other concrete material data necessary to derive constants for other concrete models. For Mode IIC (B1 > 0), the labels "effective plastic strain" and "effective stress" in the d3hsp file are incorrect. The correct labels should be "damage (lambda)" and "scale factor (eta)" as described under Mode IIC in the User's Manual. Both are dimensionless. For Mode IIB (B1 = 0), the correct labels for Cards 4 - 7 would be "effective plastic strain" and "scale factor (eta)". Only for Mode I do Cards 4 - 7 represent "pressure" and "yield (or effective) stress". Mat_72r3 also has an 'easy input' option which is based on the unconfined compressive strength (recommended if you know nothing else about the concrete). *Mat_72 says it's appropriate for impulsive loadings so I would assume *mat_72r3 would also fall under this heading. Regarding fc' and *mat_005: For *mat_005, plot sigy vs pressure using the relationship for yield stress given in the User's Manual. On the same plot, add a straight line with a slope of 3 which passes thru the origin. This straight line represents the load path of an unconfined compressive test. Where the straight line and yield curve intersect identifies fc'. fc' alone is not sufficient to provide input for *mat_005. *Mat_016 might be a better choice if that's all you have. A complete input deck for a notched beam simulation (using solid elements) using *mat_096 is available in http://ftp.lstc.com/anonymous/outgoing/support/FAQ_kw/concrete/notchbeam.mat96.k.gz *Mat_096 is probably NOT a good general purpose concrete model. *Mat_084/mat_085 can produce a binary database that permits the user to visual cracking of the material using LS-Prepost. See also http://ftp.lstc.com/anonymous/outgoing/support/FAQ_kw/concrete/mat84_winfrith *Mat_25 and *mat_145 are geologic cap models. Mat_145 is superior in the sense that the geologic cap is numerically smooth/continuous. Mat 145 elements are indeed eroded when the damage parameter exceeds 0.99; unity is the maximum value. The damage may be turned off by setting Afit=Cfit=Efit=1.0 AND Bfit=Dfit=Ffit=0.0, as explained in the LS-DYNA User Manual. The user can set FAILFG=0 (Card 5) and failed elements will not be removed. Although this is intended for SPH & ALE, it seems to fit the user's need; the elements will remain, but their strength will be zero. I plan to add a feature where the failed strength is not zero, but that awaits funding. (Len 6/6/03) The so-called Release III of the Karagozian and Case (K&C) Concrete Model is *mat_072r3 or *mat_concrete_damage_rel3. It has a model parameter generation capability based on the unconfined compressive strength of the concrete. *** NEW *** *mat_273/*mat_cdpm/*mat_concrete_damage_plastic_model is available in version R7.0.0. This model is aimed at simulations where failure of concrete structures subjected to dynamic loadings is sought. The model is based on effective stress plasticity and has a damage model based on both plastic and elastic strain measures. Implemented for solids only but both for explicit and implicit simulations. Using an implicit solution when damage is activated may trigger a slow convergence. IMFLAG = 4 or 5 can be useful. Implemented by to at Dynamore. _______________________________________________________________ RE: failure and erosion Concrete may fail in tension (spalling, cracking). A pressure or stress cutoff built into most concrete material models may be sufficient to account for tensile effects without deleting the element. When the concrete fails in compression, it may be desireable to remove the failed elements from the simulation. In v. 971, two compression-based failure criteria have been added to *mat_add_erosion. These criteria are input in the 3rd and 4th fields of Card 1: 3rd field: PMAX (max pressure) (fails if p > PMAX) 4th field: EPSP3 (fails if min prin strain < EPSP3) Card 2 of *mat_add_erosion contains primarily tension-based failure criteria. If an element meets any of the criteria specified with mat_add_erosion, the element will be deleted. ________________________________________________________________ RE: reinforcement of concrete (rebar) Materials 16, 72, 96, and 84 include the option of considering reinforcement (rebar) in a smeared fashion. Alternately, reinforcement can be modeled in a discrete manner using beam elements. This beams may be (a) merged to the solid concrete elements (shared nodes), or (b) tied to the concrete elements using 1-D contact (*contact_1D) which can account for bond slip, or (c) may be coupled to the concrete elements via *constrained_lagrange_in_solid (CTYPE=2), *ale_coupling_nodal_constraint, or *constrained_beam_in_solid (includes option for beam to slip in the longitudinal direction). *constrained_beam_in_solid (CBIS) was introduced in R7.1.2 but subsequent bugfix(es) are included only in version dev r94935 (Hao, 5/12/15). Furthermore, the capability was extended to beam types 2-6 at dev r98943. Roger checked in the implicit implementation for *CBIS in r99215. Please note that for both keywords (*CBIS and *CLIS ctype=2), implicit only supports the NQUAD=0 case. There's an article "An Introduction to *CONSTRAINED_BEAM_IN_SOLID" by Hao Chen in the FEA Information newsletter sent out on 10/26/16. Examples of rebar coupled to concrete are http://ftp.lstc.com/anonymous/outgoing/support/FAQ_kw/concrete/coupling_by_clis.k.gz and http://ftp.lstc.com/anonymous/outgoing/support/FAQ_kw/concrete/coupling_by_cbis.k.gz An example which uses *contact_1D to constrain a line of rebar nodes to a line of concrete nodes is http://ftp.lstc.com/anonymous/outgoing/support/FAQ_kw/concrete/rebar.1dcontact.rev1.k.gz See also pullbar.k for a simple example of 1D contact that verifies the theory. That theory is briefly discussed in the text file contact.1d. _________________________________________ RE: implicit analysis: An LS-DYNA distributor in the UK, expert in civil engineering applications of LS-DYNA, kindly offered these comments several years ago... "As far as I know, nobody here has attempted nonlinear analysis of concrete with implicit LS-DYNA. My understanding is that other implicit codes (e.g. ABAQUS, ADINA, etc) are OK for small amounts of nonlinearity but once it comes to cracking and softening only explicit does the job properly. We have done some work with LS-DYNA implicit on building structures, but using elastic material for the concrete. For explicit analysis of concrete using solids, we often use Mat84 (Winfrith). We wrote our own material model for concrete with shells (and beams) (MAT_CONCRETE_EC2, mat_172)." The comments still ring true in my opinion. Any implicit solver has a tendency to become unstable (won't converge) when there is a high degree of material nonlinearity that includes cracking, crushing, erosion. That does not mean you can't attempt it. There is a far better chance at success if you employ implicit dynamics rather than implicit static. _____________________________________________________________________________ RE: Example input for concrete material models LSTC not take responsibility for the 'correctness' of these constants. *KEYWORD *TITLE 4 concrete material models in English units $ $ Note that mats 16, 72r3. and 159 offer shortcut input methods whereby essentially $ only the density and unconfined compressive strength need be provided by the user. $ $ User is responsible for validity of material input. $ $ Psi, inch, sec, lbf-s2/inch4, lbf $ *MAT_SOIL_AND_FOAM $ pedigree of this mat_005 input is unknown; I think fc' is sigy when $ p=sigy/3. Plugging a0,a1,a2 into expression for sigy, fc' = sigy = 6956 psi $ assuming my hand calcs are correct. jd $# mid ro g bulk a0 a1 a2 pc 5 2.2000E-4 7.8800E+5 6.0000E+6 2.4390E+6 6025.0000 -0.051900-300.00000 $# vcr ref 0.000 0.000 $# eps1 eps2 eps3 eps4 eps5 eps6 eps7 eps8 0.000 0.020000 0.037700 0.041800 0.051300 0.100000 0.500000 0.000 $# eps9 eps10 0.000 0.000 $# p1 p2 p3 p4 p5 p6 p7 p8 0.000 21000.000 34800.000 45000.000 58000.000 1.2500E+5 9.4450E+5 0.000 $# p9 p10 0.000 0.000 *MAT_PSEUDO_TENSOR $ Mode II Concrete Model Option with automatic EOS $# mid ro g pr 16 2.2000E-4 0.000 0.200000 $# sigf a0 a1 a2 a0f a1f b1 per 5000.0000 -1.000000 0.000 0.000 0.000 0.000 0.000 0.000 $# er prr sigy etan lcp lcr 3.0000E+7 0.300000 60000.000 1.0E+5 0 0 $# x1 x2 x3 x4 x5 x6 x7 x8 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 $# x9 x10 x11 x12 x13 x14 x15 x16 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 $# ys1 ys2 ys3 ys4 ys5 ys6 ys7 ys8 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 $# ys9 ys10 ys11 ys12 ys13 ys14 ys15 ys16 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 *MAT_CONCRETE_DAMAGE_REL3 $ see rate effects curve 723 below $# mid ro pr 72R3 2.2000E-4 0.200 $# ft a0 a1 a2 b1 omega a1f 0.000 -5000.000 0.000 0.000 0.000 0.000 0.000 $# slambda nout edrop rsize ucf lcrate locwidth npts 0.000 0.000 0.000 1.000000 1.000000 723 0.000 0.000 $# lambda1 lambda2 lambda3 lambda4 lambda5 lambda6 lambda7 lambda8 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 $#lambda09 lambda10 lambda11 lambda12 lambda13 b3 a0y a1y 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 $# eta1 eta2 eta3 eta4 eta5 eta6 eta7 eta8 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 $# eta09 eta10 eta11 eta12 eta13 b2 a2f a2y 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 *MAT_CSCM_CONCRETE $# mid ro nplot incre irate erode recov itretrc 159 2.2000E-4 1 0.000 0 1.000000 0.000 0 $# pred 0.000 $ f'c in psi,,unit system $# fpc dagg units 5000.0 0.000 3 *DEFINE_CURVE $# lcid sidr sfa sfo offa offo dattyp 723 0 1000.0000 0.000 0.000 0.000 0 $# a1 o1 -30.000000 9.700000 -0.300000 9.700000 -0.100000 6.720000 -0.030000 4.500000 -0.010000 3.120000 -0.003000 2.090000 -0.001000 1.450000 -1.000000e-04 1.360000 -1.000000e-05 1.280000 -1.000000e-06 1.200000 -1.000000e-07 1.130000 -1.000000e-08 1.060000 0.000 1.000000 3.000000e-08 1.000000 1.000000e-07 1.030000 1.000000e-06 1.080000 1.000000e-05 1.140000 1.000000e-04 1.200000 0.001000 1.260000 0.003000 1.290000 0.010000 1.330000 0.030000 1.360000 0.100000 2.040000 0.300000 2.940000 30.000000 2.940000 *END $ Here's an example input that requires an equation-of-state (EOS) $ PEDIGREE UNKNOWN! *MAT_PSEUDO_TENSOR 3,2.247E-04,,0.220 500.,1.250E+03,0.333,6.667E-05,500.,1.50,1.25,0.770 3.000E+07,0.200,6.000E+04,4.031E+06,0.000E+00,0.000E+00 0.000E+00,8.620E-06,2.150E-05,3.140E-05,3.950E-04,5.170E-04,6.380E-04,7.980E-04 9.670E-04,1.410E-03,1.970E-03,2.590E-03,3.270E-03,4.000E-03,4.790E-03,0.909 0.309,0.543,0.840,0.975,1.00,0.790,0.630,0.469 0.383,0.247,0.173,0.136,0.114,8.600E-02,5.600E-02,0.000E+00 *EOS_TABULATED_COMPACTION 3,0.000E+00,0.000E+00,1.00 0.000000000E+00,-4.000000190E-03,-5.499999970E-03,-1.360000018E-02,-2.019999921E-02 -3.559999913E-02,-4.289999977E-02,-5.189999938E-02,-6.190000102E-02,-7.530000061E-02 0.000000000E+00,7250.00000,9425.00000,14065.0000,18415.0000 26100.0000,31900.0000,34800.0000,44950.0000,58000.0000 0.000000000E+00,0.000000000E+00,0.000000000E+00,0.000000000E+00,0.000000000E+00 0.000000000E+00,0.000000000E+00,0.000000000E+00,0.000000000E+00,0.000000000E+00 2550000.00,2550000.00,2550000.00,2550000.00,3340000.00 4280000.00,5220000.00,6210000.00,7500000.00,7500000.00