Damage Mechanics - Science topic

check your file type of that individual part and change accordingly.

To model a sandwich structure in Abaqus, start by defining the material properties for the face sheets (CFRC) and the honeycomb core. For the face sheets, use an orthotropic elastic material or a damage-based composite material model, specifying Young's moduli, Poisson's ratios, shear moduli, and strength values. For the honeycomb core, define orthotropic properties, including E1,E2,E3E_1, E_2, E_3E1​,E2​,E3​, ν12,ν13,ν23\nu_{12}, \nu_{13}, \nu_{23}ν12​,ν13​,ν23​, G12,G13,G23G_{12}, G_{13}, G_{23}G12​,G13​,G23​, and the compression and shear strengths in the W and L directions.

Next, create the tapered cylindrical shell geometry using either 3D shell elements for efficiency or solid elements for higher fidelity. Define a composite layup using the Composite Layup Editor, treating the core as a thick ply between the face sheets and specifying the layup sequence as [45, -45, Honeycomb, -45, 45]. Assign the composite shell section to the geometry, or, if using solid elements, partition the geometry into regions for the face sheets and core and assign material properties to each region.

For interactions between layers, use tie constraints or surface-based cohesive behavior to bond the face sheets and the core. To investigate the residual strength after impact, perform an explicit dynamic analysis to model the impact event, followed by a static analysis on the damaged structure under the desired loading conditions.

Mesh the model with a fine mesh near the impact zone and coarser mesh elsewhere, using shell elements (S4R) for the face sheets and solid elements (C3D8R) for the core if modeled explicitly. Enable failure criteria for both the face sheets (using models like Hashin or Puck) and the core (with shear or compression failure criteria). Finally, analyze the results to evaluate impact-induced damage and residual strength. If computational resources allow, consider modeling the core explicitly rather than as a ply for greater accuracy. Still for in depth please share the model??

There are vaious NDT methods, supported by analysis, to support estimation of design life and residual life of a pressure vessel. This is a specialized area, and oftern done in detail by nuclear industry and other process industry dealing with pressure vessels containing toxic chemicals, petroleum products, etc.

I asked this question for PyANSYS team also. Here is their reply:

This was a mistake from my side as I just started using PyANSYS. Functionality I was seeking is related to the PyMAPDL Reader, not PyMAPDL itself.

Regards

To define debonding at the matrix-particle interface using the cohesive surface method in Abaqus, you can follow these steps:

Keep in mind that the cohesive surface method is just one approach to modeling debonding in particulate composites. Other methods, such as the cohesive element method, may also be used depending on the specific needs of your simulation.

Miguel Angel Valdivia-Camacho Thank you! Add to your reply, the link should be

From my previous experience, I concluded that Abaqus standard doesn't truly delete elements, which means the element always appears in the formulation, but explicit do.

I am using explicit Hashin to simulate carbon/epoxy failure. Sometimes the elements remove correctly, but sometimes they don't, even if they are badly deformed with at least one of the criteria reaching 1 already. (I tried different energy degradation criteria.) Still trying to figure out why.

Here is the latest published Modelling Guide for Timber Structures developed by more than 100 global experts: https://web.fpinnovations.ca/modelling/. Chaper 4.1 discusses the constitute models for wood-based products and key modelling considerations.

Independent of which material model you take, you need to provide information on the hardening behavior in the post-necking regime in order to be able to simulate this region. Getting this information for a real material from experiments is not trivial. Though, there are established methods to extract true-stress vs. true strain data from the post-necking regime, e.g. using Bridgeman's correction or doing an inverse parameter identification (the latter is necessary definiely in case of complex stress states). A pragmatic approach is that you take the true pre-necking data and extrapolate them to the post-necking regime using an analytical law like a power law or a Voce law. Subsequently, the analytical law can be either rasterized to obtain the tabulated data for the built-in plasticity models of Abaqus or you implement it as a user-defined hardening law using the UHARD routine. Our GTN implementation (open source at https://tu-freiberg.de/nonlocalGTN ) supports power-law hardening by default.

As I just figured out, it's basically the strain energy of a body divided by its volume.

the following link gives a good explanation:

Strain Energy Density: What, Why and How? | OnScale

Use subroutines to delete elements (material points).

Maybe you can find the definition and example here:

Bests

Hi Torkan Shafighfard

Failure criteria, e.g. Maximum Stress alluded to earlier, require the tensile and compressive strengths of the lamina in each normal direction, i.e. 1, 2, 3, and shear directions, 12, 13, 23 (for plane stress, often assumed for laminate analysis, the coordinates 1, 2, 12 will suffice). These strengths be determined by straightforward tensile, compressive and shear experiments or from micromechanics (i.e. calculated based on strengths/properties of fibres and matrix and their associated volume fractions). If determining the lamina strengths by experiment we do not need to know the individual constituent strengths - we treat the material as homogeneous at the lamina level.

When analysing damage at the micro-level you will need to know the properties of the individual constituents.

Hi,

About your second question, I think the attached link can really help you. In this link, the procedure of doing fatigue analysis is completely explained and as you can see there, it states that: "It is impractical and computationally expensive to perform a cycle-by-cycle simulation for a low-cycle fatigue analysis; Instead, to accelerate the low-cycle fatigue analysis, each increment extrapolates the current damaged state in the bulk material forward over many cycles to a new damaged state after the current loading cycle is stabilized." So for the very first cycles, specially the 1st one, analysis is performed in more detail and then the next cycles will be analyzed by extrapolation based on the previous cycles. That's why the number of iterations in the very first cycles are greater than the other ones.

hope it would be helpful.

regard

DAMASK 3.0 (currently in alpha status) has images on dockerhub:

Note that these are for the grid/spectral solver and pre- and post-processing only

Linan Qiao Wow, 8 years later.... :D

Did you finally make progress with the material model of Polymer Concrete in Abakus? How did you model the plastic behavior as well as the softening behavior after reaching the max load? I would appreciate if you shared that information with me!

Dr. Zhu:

In terms of their potency, for a crack of a given length, a surface crack is at least twice as dangerous as an internal crack (Remember the Griffith equation for the ideal fracture strength where the crack length a is the total length of an edge crack and half the length of an internal crack. Also, for an infinite sheet in far-field tension, for example, the stress intensity for an edge crack is 1.12 times larger than that for an equivalent sub-surface crack which is twice as long). Because of this, fatigue cracks invariably initiate at the surface. If for some reason there are no surface cracks or they are suppressed, fatigue cracks will be driven to initiate in the interior of the sample. For the reasons described above, for a given sized crack, the stress intensity at the tip of a sub-surface crack will be more than half that for an equivalent-sized surface crack; moreover, as it extends below the surface it will be essentially growing in a vacuum and thus be free from the corrosion-assisted fatigue which may occur for the growth of cracks at the surface. Accordingly, for an equivalent distribution of flaw sizes in a sample, the fatigue life where the crack initiates sub-surface will be far in excess of that where the cracks initiate at the surface.

The question that then arises is how to dramatically extend the fatigue lifetime by causing initiation to occur below the surface, as there are data that clearly show the duality of fatigue strengths for surface and sub-surface initiation of cracks. The best way to do this is of course to put the surface in compression which is achieved industrially by techniques as shot-peening, laser shock peening, cold rolling or burnishing, case hardening, nitriding, induction hardening, etc. Many of these surface treatments also induce a work-hardening layer at the surface, which is significantly more stable at higher temperatures where the residual compressive stresses can relax out. Both the surface residual compression and the work-hardened surface layer can be extremely effective in suppressing surface crack initiation and forcing fatigue cracks to initiate sub-surface (such sub-surface cracks are often known as "fish eyes"), with the result that major increases in the fatigue life of the structure or component can be achieved.

ROR

Dear René Hammer ,

Thanks for your reply. Your new paper seems to be very interesting. I must read it. Thanks for providing the link.

regards,

Faezeh

Hello

try to get assistance from the ABAQUS mailing list

Before posting search the archive of the list.

There is not much traffic on this list lately, and mails are sent after a long time only.

Get the file

http://imechanica.org/files/Writing User Subroutines with ABAQUS.pdf

Good luck

Frank

I won't add much to what Anas said except, pay attention to loading/boundary conditions. You acquire material properties from suppliers.

Thanks Liu Tengfei! Yes, the mechanical properties seems difficult to find.

Avner Shmuel can you please tell me where you are with your implementation, I can help you from there on. Send me a message so that we can talk in detail.

Hi Qitao,

This is a problem of unstable crack propagation (i.e., fast crack propagation). In the quasi-static CZM analysis, the gradual CZM model should be established in the crack propagation process from initiation to arrest. We have a paper published recently for you to refer if you contact to me.

Looking forward we may to more communication.

I think you basically answer your question yourself. Thin-ply composites are stronger but also more brittle than standard ply composites. So yes, once damage has initiated the toughness is lower.

You did not tell in which software!

Every software has a somewhere to choice responses

I tried the simple crack band model regularization which adjusts the dissipated energy with the mesh size. In my damage model the softening modulus changes based on mesh size. I have not introduced any weak element to trigger localization.

Maybe you can focus on the shape of the cavities.

Assume under uniaxial creep testing:

If the plastic hole growth takes control, then the cavities should be more like a standing needles which is perpendicular to the GB. Otherwise if the vacancy diffusion mechanism controls, the cavities should be like cracks that is parallel to the GB.

But to remember, when the constrained diffusional growth mechanism takes control, then the cavities growth rate actually depends on the slow creep deformation on the neighbourhoods matrix. However, the cavities look like the crack-type, since the vacancy diffusion contributes most of the growth but it is constrained by surrounding slow creep deformation rate. This is the most common type of creep cavitation.

Aamir Sultan

see these articles :

Part I:

Part II:

Dear Dhanraj Rajaraman

Please find the attached file.

Dear Seyed Reza Jafari ,

I defined the damage on the base of experimental measurements using a multi-linear-elastic material model.

Sir, are you able to include pinching in your model ? if so then please guide me how to include pinching in the hysteretic model.?

will be very thankful.

Mehdi Moghaddasi

It is a very interested question

Dear Victor,

I think these books can help you understand well.

- Book, Numerical Simulation of Mechanical Behavior of Composite Materials.

Author: Sergio Oller.

- Book, Fractura mecánica. Un enfoque global (in Spanish) .

Author: Sergio Oller.

I hope it is useful.

Best regards,

Jorge

Hi, dear Behrooz,

for me changing element size and trying to place crack within the elements worked out.

Here's a working, official Simulia version of the VUMAT you're trying to utilize: https://pastebin.com/1Mx7WYed

The JC model is supposed to provide you with the effective plastic strain rate. Its implementation within a explicit FE solver is quite straight forward. You should use the following steps:

STEP 1: Calculate the equivalent stress s_eq at beginnning of the time increment with the yield criterion of your choice (e.g. von Mises, Tresca...)

s_eq=f(s)

STEP 2: Calculate the plastic flow direction n according to:

n_ij=ds_eq / ds_ij

STEP 3: Calculate the equivalent plastic strain rate depeq/dt from the JC constitutive equation

de_p,eq/dt = f(s_eq, T, e_p,eq)

STEP 4: Calculate the plastic strain rate tensor

de_p,ij/dt=de_p,eq/dt*n_ij

STEP 5: Calculate the plastic strain tensor and equivalent plastic strain at the end of the time increment

e_p,ij=e_p,ij+de_p,ij/dt*dt

e_p,eq=e_p,eq+de_p,eq/dt*dt

STEP 6: Calculate the damage indicator at the end of the time increment

D=e_p,eq/e_f

STEP 7: Calculate the stress state at the end of the time increment

s_ij=s_ij+C_ijkl (de_kl -de_pkl)

where C is the fourth rank stiffness tensor

Answering w/o justification or complementing on questions are not welcome on an academic website. Here is not Linkedin or Facebook. Getting back to Shima's question. Usually Taguchi DOE is used in cases where you want to perform as minimum as possible number of iterations (or Cases). In my opinion, full factorial DOE provides the best accuracy, however, if you are limited in terms of time or expense associated with your analysis, then you can use other methods with less accuracy.

good qustion

Hi,

There are numerous articles available where 2D Puck's theory and 3D have been implemented. For instance in this article, 2D version has been implemented:

BR

@ Hamid, if I am correct you suggest to use Engineering stress and Engineering strain values and find the point where yielding occurs. Then, use the simple formula to calculate 'dc' i.e. the damage parameter in compression. How to obtain 'dt' ?

Please, see:

It depends on how complex the material behavior is, as well as on limitations associated with built-in constitutive laws e.g. the hyperelastic laws do not allow for damage but the viscoplastic do. Therefore VUSDFLD through field variables will not help when using a built-in hyper(visco)elastic law and in this case VUMAT/UMAT is necessary. However debugging the VUMAT routines through simple models i.e. single elements does not guarantee reasonable behavior in the real application model especially when large strains & element deletion take place. You have to account for sufficient energy dissipation to avoid such problems. The following paper gives a practical-easy solution for modelling large strain fracture for non-linear materials in ABAQUS:

SKAMNIOTIS, C. G.; ELLIOTT, M.; CHARALAMBIDES, M. N. On modeling the large strain fracture behaviour of soft viscous foods. Physics of Fluids, 2017, 29.12: 121610.

Yes but no any source for H2Sin that process, we need to know about hydrogen embrittlement in duplex ,what is possible sources of H2 from manufacturer it could be ??

Hello everyone! I'm working on a user subroutine USDFLD to run with my input fine! The subroutine is running well and S11, S33, S12, S23, S13 are correct but the stress values of S22 from the analysis is not correct from SDVs! I don't know where I am making the mistake! Can anyone please help me with this.

Danilo indeed lists the correct relationships to be used in Abaqus. To understand the derivation, it helps to look at when the Paris law is linear (e.g., on log scale). e.g.

log(da/dN) = m log(K) +log (C)

And so, "m" is the slope of the linear relationship, and "C" is the intercept.

As mentioned, the following relation hold:

G=K²/E

or

K = (GE)^1/2

Plugging this into the first equation:

log(da/dN) = m log( (GE)^1/2 ) + log C

log(da/dN) = (m/2 log (G) )+ (m/2 log E) + log C

This is still a linear relationship, now the slope is m/2, and the intercept is (m/2 log E)+log C. Bringing this back into the original scale (e.g. take exp()) You'll indeed see:

da/dN= (CE^m/2)(G^m/2)

or

C3=CE^m/2

C4=m/2

Of course not. ANSYS will only instantaneously degrade the stiffness for the specific failure model.

So to answer your question again, : When the matrix fails according to Puck’s failure criterion (one of the modes A,B,C) will damaged elasticity matrix have a reduced stiffness in the fiber’s direction?

The answer is NO. Ansys will only degrade the stiffness for the specific failure mode.

Hi,

I am also looking for answer for the same question. If you got any answers, can you please share with me. I also certain doubts in Stress-strain hysteresis curve, so can you help me with this. If so, please write me an e-mail at jayantshanmugam@gmail.com

Thanks.

With Regards

Jayant

I had the same issue and found out the problem to be I need to define a section for the new part after merging. As you know after merging parts you have a new part that you need to assign a section for.

I hope the answer is helpful.

yes, it is the most useful one. Recently Ranjbaran &Ranjbaran and their research team proposed the so called "state based philosophy (SBP)" where every phenomenon may be treated based on exact reasoning and mathematical logic. The SBP applied to the fracture mechanics and all parameters of fracture mechanics are explicitly and exactly determined. The new fracture mechanics, is in close agreement with the test results. It is different from the classical fracture mechanics (empirical and/or finite element based) in the sens that it is based on exact mathematical logic. Therefore all ambiguities in the classical fracture mechanics (singularity, the unknown domain of crack effect, the strain released energy, ant etc.) are no longer present. The new fracture mechanics led us to solve other damage problems, e.g. buckling, yielding, and etc., with the same concept.

Ali,

I had experience in different types on Damage for a number of materials in LS-Dyna including Johnson Cook. However not in the machining simulation area.

An example on some of the different types on element formulations that can be used in a numerical machining simulation (From the reference above) are as per below (Lagranian, Eulerian, ALE and SPH):

An in depth discussion on these methods in LSDyna (Lagranian, Eulerian, ALE and SPH) that can be used for both solids and fluids  are discussed in:

3.3.3. Numerical Methods

From first principles my guess is that explicit Under Integrated  Lagranian element approach would be a good starting point with auto mass scaling to deal with distorted elements and a finely tuning of the element deletion criterion (may only be needed near to the cutting surface?). However SPH may be more robust (Easy to use) but could be less accurate?

The Johnson-Cook material model with failure is a classic model, and you will find heaps of test materials on the web. If you can not find the specific one you may need to test it?

Regards,

David

Start with a simple model that you can corroborate with data that everyone agrees on, and then run it using different erosion models.  At this stage, I wouldn't worry about -having- to use a particular one (e.g., Johnson-Cook), but I would want to find one that gives good results and that is workable / solveable.  If you can solve that simple model with J-C and the results agree with the agreed-upon data, then you'll have your JC constants from that exercise.  If you need additional help in deriving these constants, then I'd contact your tech-support person from LSTC, if that is available to you (sometimes there is only one person at the university that the software provider will work with - you'll have to ask around if that is true for your university).  Hope that helps. 

another nice link:

You can see the help document as followed:

Abaqus Example Problems Manual 1.4.7

Inputfiles dcb_vcct_fatigue_2d.inp

The help is a good tool for studying.

 

hi

i cant find the AIR material in your model.

after define the model and property of material,assemble and set position, in interaction all model must be coupled .

so the in predefined, you must be material assignment and initial stress define.

...

i think you probably can attempt to use different damage evolution in concrete constitutive model or create bond-slip between longitudinal rebar and concrete matrix. this will result in more remarkable unloading stiffness and larger deformation.best wishes!

Thank you very much everyone for the valuable comments. My error was in the mesh because I have missed to assign the element type. Now my model works fine. Thanks again. 

Check out this paper, you may have some idea about smeared cracking and damaged plasticity models. 

Hi, have you tried to hide rebar from current viewport in display group menu?

I think you have to choose a trial value for the maximum principal strain and observe response it gave until you get a failure load close to what you are looking for. Secondly the variation should equally involve the other damage criteria like the fracture energy,depending in what your criteria are. 

You can also read more about it in the Abaqus documentation. 

Thank you and good luck. 

Hi there,

I couldn't clearly understand your model, but  if I am correct, It seems like a frictional tool tip goes (or forcing the material) through the work piece, (viz. punching ?).  

If you couldn't get the elements deleted while using a material failure based erosion parameter, then I guess that you may try assigning an erosion parameter for each material such as (geometrical) strain limit. It should be a solver control parameter. I am not experienced with metals but the 1.5  value (or between 0.5 - 2) generally works for most materials for this type of loading. Mesh density/quality would affect your results since element erosion can affect the physical response. Therefore, you would like to have an experimental result of the specific application to compare the simulation in terms of some parameter correlation to decide  which value works better.

Hope it helps. 

Asim

Hi Sylvain,

I think that you can find a solution to your problem on the Abaqus verification manual at the following link:

Looking for these input files:

cdpu_cps4r_sx_s.inp       Abaqus/Standard base analysis.

cdpu_cps4r_sx_x.inp       Abaqus/Explicit import analysis from cdpu_cps4r_sx_s.inp.

Federico

Stresses are element quantities and are always calculated at integration/section point quantities ins shells...Your question is not clear as to what exactly you want.

in riks method LPF will never become constant as the response moves from pre- to at- to post buckling regimes

We are working on this problem too! Now we have encountered 2 major problems.

A: Fatigue Crack Growth(FCG) Rate

Crack growth under compression may cause a lot of problems. Since the crack closure and the crack face friction take place during cyclic loading. The crack driving force may be changed by the (tangential) friction force, and fractography may also be destroyed for same reason, which means we could barely find any clue on fractography under compression.

B: Fatigue Crack Growth Path

Opening crack grows in its local Mode I direction, based on LEFM (also MTS criterion), the maximum tensile stress plane. However, tensile stress cannot be find under compression at the crack tip, which means a totally different propagation mechanism. We are trying to figure out what the FCG mechanism is and build up some relationship between the mechanism and FCG path.

Best wishes! Looking forward your solution!

Hi Vahid,

Check this article:

Hi.

You can use this process:

Step --> Output --> History output request --> Create --> Domain: Crack: (select your predefined crack) --> Frequency: every n increments --> n: 1 --> number of contours: (type your desire number, I recommend 10) --> type: Stress intensity factor --> crack initiation criterion: (select which one you want, I recommend KII = 0) --> save

After completing Job, in result: history output section, you can select stress intensity factor to plot.

Hi,

in VUMAT as well as other material subroutines, the strains (not displacements) resp. strain increments are input arguments to the stress update which cannot be modified.

You may try built-in Abaqus features instead, allowing the prescription of initial strain or displacement fields.

Regards

Thanks Dr

Optonor have a new,  a specialized shearography technology which can be used. 

Refer to fracture mechanics...fatigue analysis of crack initiation from maximum stress concentration at crack tip, then the crack propagates with decreasing of stress with increasing of crack length relations with stress intensity factor.

Dear  Ashraf:

you can get the overall damage per element (including both tension and compression damage) by select stiffness degradation output fro field output (SDEG) the per cent is included automatically in the legend (if not check the setting).

note: tension and compression damage is one way variables (no recovery)

stiffness degradation output consider the stiffness recovery for more details please check abaqus material manual

 Best of luck