Science topic

Abaqus - Science topic

Explore the latest questions and answers in Abaqus, and find Abaqus experts.
Questions related to Abaqus
• asked a question related to Abaqus
Question
My attempt was that I created 2 Reference Points at the centers of the two base of the cylinder and made a wire between them. Using that wire, I made a planar connector. But when I Provide displacement BC to the cylinder, it also moves above the plane.
I expected that the result would be something like the cylinder slides on the plane and due to the upward displacement, a stress is generated.
I believe there is something wrong with the way I made the connector. (I am very new to this software)
I suggest you to use zero thickness interface element between the cylinder and thin sheet. In abaqus there is a surface to surface contact element available, use this for simulating the contact between two dissimilar materials.
• asked a question related to Abaqus
Question
I am trying to model the combined effect of a static and dynamic load on a cantilever beam. I have calculated the dynamic response as follows
1) Natural frequency and mode shape extraction (modes 1-5), see image
2) Steady-state modal analysis gives the steady-state response across a range of frequencies. I am looking for deflection with direct model damping of 2%. If I am not mistaken, this form of analysis require a load independent of time and essential applies harmonic loading to each of the natural frequencies and calculates the response at each frequency.
I am unable to apply a static load at this stage (steady-state modal dynamics) to observe the combined effect. See image for the deflection as a function of frequency. This is precisely the plot I want if I was just considering dynamic response. But surely, this would change significantly with a static load.
I tried setting up a static load in a step prior to the frequency and steady-state modal dynamics step but that was of no use.
Any tips on how to do this.
FYI, I am actually trying to model a chassis experiencing cyclic and static loading from an engine and am using this to fully understand the problem.
Gabriel Zenkner, I had the same issue and I think I have found the solution. Create 2 steps, the static and the dynamic, and submit the analysis. Then, once you have the solutions to both problems, go to Visualization Module -> Tools -> Create Field Output from Fields. There, you can select Fields from each Step and Frame Increment. Once you pick one, it automatically adds the Field Variable to the big white upper box where you can write whatever expression you want. Also, to the right, you have a list of available operators. In the following example, I wanted to obtain the combined stress state due to static and harmonic loads, therefore: s3f1_S + complexReal(s2f2_S) (stress tensor due to static loads on frame 1 + real part of stress tensor due to harmonic loads on frame 2). The complexReal() operator is necessary since it appears that the harmonic stress tensor in stored in complex form.
Even though I selected to compute the real response only in the steady-state dynamic step, some errors show up when I tried to do the sum without the complexReal() operator.
• asked a question related to Abaqus
Question
I am trying to do a simulation of a model (with static general and coupled temp-displacement steps) using a user subroutine (UMAT).
But when I run the job, this fails because it needs too many increments to complete the step.
Attached are some of the warnings I get.
Would anyone know how to fix this problem?
Unfortunately I tried to modify the maximum number of increments and the minimum increment size of each but it didn't work.
I will try to look in the outputs database of the visualization module to see the locations of the warnings.
Regards
• asked a question related to Abaqus
Question
Dear All,
when we model the bond-slip behavior between reinforcement and concrete in ABAQUS, should the reinforcement always be modeled as a 3D solid element.?
Is it not possible if with Wire element type, because in my model used Wire element.? if possible with Wire element type, how do I model it in ABAQUS, and what contact type should I choose.?
Hi Puji Kurniawan , I just want suggest my friend after I saw your profile. Please talk to Indra Komara( my best friend) from Bandung. He'll absolutely help you.....He's just "orang pintar ". Telepon saja loh!
• asked a question related to Abaqus
Question
Hello all,
I am relatively new in area of simulation and I have this basic doubt. I want to know the meaning of time when we are running static simulation. Static means load/BC don't change with time but we give values on time scale.(For abaqus, we define step with total time, initial increment, min/max. incerement) What does it actaully mean? Is it related to - how solver deals with problem and runs the simulation in background?
I am expecting your thoughts/inputs on this topic.
Regards,
Rohit
By default, your load will increase linearly as Magnitude*t/Time period, (0<=t<=Time period), where t ("time") and Time period are not necessarily measured in seconds. If you want the load to be constant after t>=10, define a piece-wise linear amplitude.
• asked a question related to Abaqus
Question
Hi,
I want to connect two nodes by using a multipoint constraint element MPC184 ( rigid beam / rigid link)
at first, i tried connecting both nodes by a line and then mesh the line and give it MPC184 element attributes but I cannot create a line from node to node on ANSYS. Only from keypoint to keypoint.
The attached image shows what i mean.
how to fix that?
Hello,
As you work with a script in M-APDL environnement, I think it's quite easy to fix your problem :
* First, define an element type associated to the MPC184 element -> ET,10,MPC184 for example, in order to associated the MPC184 element to the element type n°10
* Then, adjust the key options in order to choose your MPC behaviour -> KEYOPT,10,1,3 for example to say that the keyopt n°1 for the element type n°10 is 3 / if the element type n°10 is associated to the MPC184 element, that means that you ask for a slider element behaviour (help MPC184 command line for help on MPC184 element and keyopts)
4/ as NODE1 and NODE2 exist, the sequence to define an element between NODE1 and NODE2 is (\$ for new line character) : TYPE,10 \$ E,NODE1,NODE2
(active the n°10 element type,which is already associated to the MPC184 element, with the appropriate keyopts, and then define an 2-nodes element between NODE1 and NODE2.
MPC184 is not associated to any geometric entity, that's why you can't mesh a line with a MPC184 element. You have to define it directly with the nodes.
• asked a question related to Abaqus
Question
Hello,
I have been modelling using concrete damage plasticity material in Abaqus 2021. After the analysis, I got DamageT values greater than 0.9, the max damage parameter. I've tried to restate the damage and strain columns together, but it doesn't solve the problem. Is there any other solutions?
I would be grateful if anybody could help me.
Thank you
Where do you see these values greater 0.9? If this is only in a contour plot, be aware that these plots extrapolate from integration points using a method that works well only for linear elastic behaviour.
Use tools->query to check actual values at integration points.
It may also be that the damage variable is allowed to evolve beyond failure (I don't know how your failure model works). You could test that witha simple model with 1 element.
• asked a question related to Abaqus
Question
Hello,
I'm trying to run my Abaqus simulation using gpus. I have a PC with AMD Ryzen™ 5 2400G & Radeon™ RX Vega 11 Graphics.
Calling the gpu from CAE and the Command window doesn't work. I have found a possible solution using CUDA, which I have not tried yet since it refers to NVIDIA hardware. Other posts suggest using OpenCL but I can not find how to download it.
You should use OPEN GL for the AMD graphics card. I'm not sure AMD supports Abaqus parallel computing.
• asked a question related to Abaqus
Question
Hello, I encounter a problem on abaqus linked to convergence. I even reduced the size of the mesh and decreased the values ​​related to `` arc lengh increment '' but it's still the same, I still have convergence errors. thank you.
Salut James
Merci pour ton apport, je vais essayer avec ca.
• asked a question related to Abaqus
Question
Does anyone have tutorial for Abaqus Additive Manufacturing simulation?
I want to simulate a 3d printed item (material of item = ABS plastic) and conduct stress analysis on it. However, I hardly can find any tutorial from Youtube and Google.
Does anyone know where to search for tutorials?
This link will direct you on the usage of Abaqus in the additive manufacturing
• asked a question related to Abaqus
Question
Hi
I need to model a Polycarbonate base adherend and on top of it we Injection mould a Polyamide material(Which acts as adhesive) and sticks very well to the polycarbonate(refer attached image). So I need to model this one in Abaqus and to find its peel strength and fracture energy. My questions are
is it possible to model this type in Abaqus?
Which type of test modelling of adherend that I need to consider SLJ, DCB, ENF?
Which method need to implement (Surface based cohesive), cohesive element, XFEM?
I am very new to this field and I want to know any reference books on determing peel strength adhesive joints in Abaqus.
@Revella Nagarjuna
It is possible only if the position in electrochemical series of carbon and Amid is understood that is the bond formation link between hydrogen and nitrogen is placed on stimulation using assemble and fracture mechanics software to obtain the new molecular structure atoms are re-arranged
• asked a question related to Abaqus
Question
I am using Abaqus and in meshing, I need distance from node to node, to be constant along Z axis. This can be seen in figures attached here. Yellow dotted lines represent 0.5 mm distance which I require and black solid lines represent mesh created by Abaqus (which varies along z axis). It is required to be parallel along x axis. As it can be seen that yellow dotted line (target distance) and black, elements boundary lines, do not coincide. So my way, is not working.
Does anyone know how node to to node distance can be kept same along Z axis despite asymmetric features in part geometry? Please comment.
Can 'assign stack direction' sort this problem out? I have tried but did not help.
Thanks.
Here is, how it looks like after completion.
So to sum up, do partitioning near source of asymmetry (in this case circle), assign one element in thickness in these partitioning, and then mesh it.
Thank you all.
• asked a question related to Abaqus
Question
Hello,
I am a new user of ABAQUS/CAE and is working on ballistic impact simulation on fabric. I referred many research article for the same, but doesn't got the clear image about its simulation.
It would be very beneficiary for me if anyone can let me now how to do material modelling of a fabric or refer some good article.
Thank You
Hello sir,
Thank You for your response, these are really good inputs.
I have another question: How will the inter woven characteristics (like the weave pattern, crimp) of the fabric taken into account?
Should I model the fabric in pre-processing software like TexGen or WiseTex and then have to write a user-subroutine for the same?
• asked a question related to Abaqus
Question
Hi dear researchers
Im trying to model polymeric material, but the elastic module of that is not constant and varies with change in strain, this polymeric strip(membrane element) is embedded in soil as a reinforcement and tensile load will be applied to that, so I wanna know if for my purpose using property >> mechanical >> elastic >>Hyper elastic is suitable or not, because as I know entering stress -strain test data is possible in hyper elastic.
Im wondering if using hyper elastic requires any special conditions or not or any alternative( better) way for my purpose.
Sujeet Kumar dear sujeet
thanks for your attention, but I have doubt about that, because my purpose is to use stress-strain curve in Elastic phase Not Plastic ,I mean using plasticity option for my goal seems wrong , whats your opinion?
• asked a question related to Abaqus
Question
Dear all,
I would like to update the yield strength value in a UMAT subroutine in Abaqus. This would happen when a load is applied on an element, the deformation strain would be queried from a meta-model database, and depending on the deformation strain combinations returned by abaqus, the yield strength of the material in the function PROPS() would be updated from the database. I am very new to writing subroutines in ABAQUS. I was wondering if there is a way to implement such an algorithm to update the material properties in each time step. If so, how can I start tackling this issue?
Any help is greatly appreciated. Thanks.
the material properties in props array of a umat are constant for the whole analysis and are defined in the material card at the preprocessing stage. If I understand correctly, you have developed a user material that calculates the stress update and the material Jacobian for each increment. In that case the material properties that are needed for your calculations would be retrieved from the meta database (you need to have access of this database in the umat) and not from the props array that is passed as argument in the umat. Then you can proceed with the calculations using the correct value of your material properties.
• asked a question related to Abaqus
Question
Every time I submit the properties, Abaqus shows the same problem about a value in the Johnson-cook damage model.
Sayyed Behzad Abdellahi I have overcome the problem, thank you so much for your suggestion :)
• asked a question related to Abaqus
Question
Hello
I linked abaqus 2017 with visual studio Ultimate 2013 and intel parallel studio xe 2017 and checked abaqus verification. It all shows a correct pass
And I've coded a simple umat, but when I run a the job it shows: Problem during linking - Abaqus/Standard User Subroutines.  This error may be due to a mismatch in the Abaqus user subroutine arguments.  These arguments sometimes change from release to release, so user subroutines  used with a previous release of Abaqus may need to be adjusted
And when I run it from the command prompt it shows: LINK: Fatal error LNK1181: Cannot open the input file 'oldnames.lib'
• asked a question related to Abaqus
Question
Hello,
I am trying to study a contact behavior with Abaqus . but after my simulation I always find that the ETOTAL energy is not constant.
How can I verify that my my model is correct or no?!
I attached the energy balance's curves and ETOTAL's curve.
Regards
I attached here the evolution of the ETOTAL during the time priode=1s
This evoltion it is normal?
• asked a question related to Abaqus
Question
I would like to ask you how can I simulate wind pressure on wind turbine towers in Abaqus as in the attached picture.
I do not know if you want to apply simultaneously the given pressure in the cylinder (as in the attached diagram). Just as an idea:
1. When you generate the geometry, you can divide the cylinder in four sectors (by using cutting planes) and then, you can apply the required pressure at each sector. Tip: In case you whant to change the distribution (or parametrize your problem), if you create a Python script, it would be easier.
2. In case you can apply symmetry in the x-axis, then, you can simulate only 1/2 of your problem.
I hope my suggestion helps.
Best regards,
Ramiro
• asked a question related to Abaqus
Question
I've built simply supported beam in abaqus.
it is easy to find the deflection, but I can not find the rotation angle.
I appreciate all the comment and advice.
Looking forward to hearing back ....
Dear Roshan Kumar Malik, i am faced like your problems. Do you have any way to get the rotation?
Thanks.
• asked a question related to Abaqus
Question
I am runing an abaqus job on HPCF cluster( high performance computing facility ) using SSH .
If anyone can share a command for killing an active job. ?
Thanks
You can use "qdel jobid".
• asked a question related to Abaqus
Question
Hi
I know the procedure of modeling modal analysis to obtain natural frequencies(I mean using Linear perturbation step>>frequency and without defining any other loading....) but the problem is all examples I have seen is just for model made of just one material(most examples for cantilever beam) but my model is consist of two or more material( 2layer soil and layer of polymer between these soil layers), for obtaining natural frequencies in this case, my exact question is: I have to run modal analysis for each material separately or all together?!
I should mention that the polymeric membrane part does not have any type of support and its just in interaction with soil( normal and tangential behaviour), and if I run the modal analysis for this part alone natural frequencies for polymeric membrane is zero. while the soil is fixed at bottom and sides and by running modal analysis once for soil alone and another time soil with polymer,the natural frequencies are the same for both conditions. so i'm willing to know any advise and special point to obtain natural frequencies procedure for model consist of two or more materials in abaqus.
Interesting
• asked a question related to Abaqus
Question
Abaqus software details needed
Abdullah Yousaf Can you share your Abaqus model (.inp)?
• asked a question related to Abaqus
Question
Hello All,
I came across an additive manufacturing heat transfer FEM problem. In this problem, in each time step, a material is added to the current geometry. So, an amount of mass with a defined temperature is added to the whole geometry.
I heard about the Abaqus AM tool, but I would like to develop my own code. So I am wondering about the best method to solve this FEM? How can I define and update the stiffness matrixes? How can I reduce computational costs?
I appreciate any help or tip!
About the stiffness matrix size: you can do it, but it is just a purely numerical solution (no science behind) if you make your steps independent (I mean if you move from step n to step n+1 and between those 2 steps there is a dimension change, export results from step n and make a transfer to a new model corresponding to your step n+1 putting zeros wherever there is no data from the previous step results and add whatever you want to add in the formulation of the initial state of your new calculation, ie the n+1) however, about adding mass, yes you would be infringing the fundamental law of mass conservation if you consider your system at step n and step n+1 the same, this is why I propose to numerically separate the model into one corresponding to step n and one other to step n+1. I hope I explained well my point
• asked a question related to Abaqus
Question
hello everyone , am using VS 2013, intel parallel studio 2018 and abaqus 2018.
when I try to run the subroutine this error shows up "LINK : fatal error LNK1181: cannot open input file 'vcruntime.lib' " any idea how to fix that :)?
I'm facing the same problem. but with ibput file 'oldnames.lib'
If you have the solution, please let me know.
• asked a question related to Abaqus
Question
Hi,
I am trying to run a vumat subroutine, but there is some abaqus-fortran linking issue.
I use
abaqus 6146
Micrasoft visual studio 14.0
parallel studio XE 2016
I added these command in my .bat file:
@echo off
call "C:\Program Files (x86)\IntelSWTools\parallel_studio_xe_2016.3.059\compilers_and_libraries_2016\windows\bin\ifortvars.bat" intel64 vs2015
call "C:\Program Files (x86)\IntelSWTools\parallel_studio_xe_2016.3.059\compilers_and_libraries_2016\windows\bin\ipsxe-comp-vars.bat" intel64 vs2015
call "C:\Program Files (x86)\Microsoft Visual Studio 14.0\VC\vcvarsall.bat"
"C:\Program Files\Abaqus\6.14-6\6.14-6\code\bin\abq6146.exe" %*
but I got the error message:
ifort: command line warning #10161: unrecognized source type 'Files\Abaqus\6.14-6\6.14-6\code\include"'; object file assumed
ifort: warning #10145: no action performed for file 'Files\Abaqus\6.14-6\6.14-6\code\include"'
End Compiling Single Precision Abaqus/Explicit User Subroutines
Begin Linking Single Precision Abaqus/Explicit User Subroutines
LINK : fatal error LNK1181: cannot open input file 'oldnames.lib'
Abaqus Error: Problem during linking - Single Precision Abaqus/Explicit User Subroutines.
This error may be due to a mismatch in the Abaqus user subroutine arguments.
These arguments sometimes change from release to release, so user subroutines
used with a previous release of Abaqus may need to be adjusted.
When I checked the system information of abaqus, it showed it is linked with intel fortran compiler 16.0, but when I ran the abaqus vertify, I got the sam error message as above for the subroutine linking.
hi mr Yun Xie i'am facing the same issue (cannot open input file 'oldnames.lib') ,i'am using abaqus 2017 linked with vs ultimate 2013 and intel parallel studio xe 2017. could you help me please?
• asked a question related to Abaqus
Question
Dear All,
I just Installed Abaqus in my computer and Abaqus CAE 6.14 works fine. But, every time I submit a Job it runs for 2 seconds and then I get the following message:
The executable standard.exe aborted with system error code 1073741795. Please check the .dat, .msg, and .sta files for error messages if the files exist. If there are no error messages and you cannot resolve the problem, please run the command "abaqus job=support information=support" to report and save your system information. Use the same command to run Abaqus that you used when the problem occurred. Please contact your local Abaqus support office and send them the input file, the file support.log which you just created, the executable name, and the error code.
I looked for information already and It seems to be something related to memory allocation but I am not sure...
does anyone have ever get this error before? do you know anything about it? is there any solution??
my computer is a NEW Lenovo Laptop - AMD Ryzen 5 4600H with Radeon Graphics 3.00 GHz - 16GB Memory - 500GB SSD with windows 10.
may be you can share your model to me I will check it if I could fins any solution.
• asked a question related to Abaqus
Question
What is the shell orientation mean in ANSYS APDL?
Does it mean that if I have a shell on the XZ plane then I set the orientation as 0 degrees as shown in the pictorial view
and if I have a shell on the XY plane then I set the orientation as 90 degrees?
Or does it mean something else? Like the direction of the fibers in the material?
The material orientation is used in case the model is described by an anisotropic material. It shows the principal axis of the element (for composite materials the fibres direction as you mentioned), where in shell elements a single angle is enough to describe the local orientation of the element.
• asked a question related to Abaqus
Question
I am doing a simulation on abaqus and my job stopped in the middle of the increment, is there a way to restart the calculation or did my job stop?
If your job stops for convergence reasons,
you can see it from the .log file.
If you see 1U, 2U, 3U et cetera
on the left side,
then it means your job stopped for convergence problems,
not because the machine crashed.
Nicolò
• asked a question related to Abaqus
Question
I simulated a cylinder of length 75 and radius 2 in Abaqus using shell elements (S4R) for the cylindrical shell and solid elements (C3D10) for flanges. The connection between flanges and cylinder is done via shell-to-solid constraint. The thickness of the cylinder is varying from 0.03 at bottom to 0.013 at top. At top a concentrated lateral force of 100 is applied. Bottom edge is fixed.
I wanted to validate my results by comparing it with a beam. Thus, a beam is created of length 75. The thickness again is varying from 0.03 to 0.013. Same BC and load condtions are applied here. No influence of flanges here, just a straight beam divided into parts with different thickness. B31 elements are used.
A linear static analysis is followed. While displacements are close, I noticed something stange regarding Von Mises stresses. The location of S.Mises at both models is different. In cylindrical shell max Mises is at boundary while in beam model is above boundary. Any idea why is this happening? I was expecting also S.Mises for beam to be at boundary
THE ANSWER BY SPYROS ARVANTIS APPEARS REASONABLE.
• asked a question related to Abaqus
Question
I am trying to perform thermal analysis of dam using abaqus but i am not able to input the adiabatic heat rise in my Finite element model further for fully coupled model is there different type of element required, What should be the boundary condition, and will the input ambient temperature and adiabatic temperature rise combine o provide resultant temperature, also can i monitor crack development?
Following.
• asked a question related to Abaqus
Question
I want to simulate a wind turbine's tower model. The cylindrical part constructed using shell elements (S4R) and the flanges using solid elements(C3D10).Internal stiffneres also included using shell elemnts. Shell and flanges are connected via shell-to-solid constraints.Only geometric non linear analysis is considered with RIKS method.
However Abaqus give me the following warning:Excessive distortion at a total of 4732 integration points in solid (continuum) elements. This warning appears when initial RIKS step is 0.02 and maximum step is also 0.02.However, when I increase the initial step at 0.05 and maximum step at 0.05 I get no warning.
What is this error about? Can affect my results and why initial step has an influence? And how to solve it?
PS: I also upload my .inp file
Spyros Arvanitis This warning is due to unsuccessful attempt number1 on increment 38, when load proportionality factor increment was chosen too large (-1310.74). Correspondingly, distortional deformations took place. The attempt number 4 become successful. Of course, this was in vicinity of bifurcation point. The Riks method with larger initial increment (0.05) have passed this point more "smoothly". The choice of initial increment must not affect postbuckling results (at least if there are no neighboring postbuckling equilibrium states).
• asked a question related to Abaqus
Question
Hi, i am doing hertzian contact 2D ball-on-flat analysis with axis symmetry elements by imposing concentrated load boundary conditions.
Interactions : surface to surface with small sliding, frictionless and hard contact properties are mentioned. Finite mesh size is used at contact point as shown in Figure.
When i am validating the FEM results with analytical values, my contact pressure (CPRESS) is with in 2 percent error. However, my shear stress(S12) value has 30 percent error. Can anyone help how to resolve this issue?
Not that this solves your problem, but at MATLAB Central you'll find a script that solves the linear complementarity problem associated with the Boussinesq contact mechanics problem
h = K*p+g, h'*p=0, h,p >=0,
where the unknown h is the gap between two elastic (half-space) bodies and the unknown p is the corresponding contact pressure.
Andreas Almqvist (2021). A pivoting algorithm solving linear complementarity problems (https://www.mathworks.com/matlabcentral/fileexchange/41485-a-pivoting-algorithm-solving-linear-complementarity-problems), MATLAB Central File Exchange. Retrieved July 11, 2021.
Andreas Almqvist (2021). An LCP solution of the linear elastic contact mechanics problem (https://www.mathworks.com/matlabcentral/fileexchange/43216-an-lcp-solution-of-the-linear-elastic-contact-mechanics-problem), MATLAB Central File Exchange. Retrieved July 11, 2021.
• asked a question related to Abaqus
Question
Hi,
I am trying to understand Abaqus Benchmark example 1.19.2 - Crack proagation in a plate with a hole simulated using XFEM.
I ran the python script provided "crackprop_hole_lefm_xfem_cpe4.py" from the site: http://dsk.ippt.pan.pl/docs/abaqus/v6.13/books/bmk/default.htm?startat=ch01s19ach135.html
I want to try and output figure 1.19.2 - 4.
I have create a path --> each node starting from left to the end of the plate. Next, Tools > XYdata > Create Path > Selected defined Path (maybe issue here) > Define Field output and frame > Plot and I cannot get the same curve.
Potential Issue:
1. Crack path was not define properly.
2. Field output was defined wrongly. I'm assuming Mises?
3. Frame was defined wrongly.
I just want to learn to plot a load vs crack length curve.
Any help is greatly appreciated!
Can you explain how to get Phi from output or at least provide the Abaqus manual where it is explained.
Regards
Biswabhanu
• asked a question related to Abaqus
Question
Hi researchers
Im trying to model dense sand which has peak friction angle and constant volume friction angle . As you k now if we plot shear stress - shear displacement(strain) of dense sand till reaching specific displacement curve ascends to the peak point(peak friction angle)and after that with slight reduction becomes constant(constant volume or residual friction angle). I wanna model the behaviour of soil accurately so my purpose is to define friction angle as a function of shear displacement. I searched and understood using SUBROUTINE (specially USDFLD) may help me but I have not work with subroutines at all and this is my first time however I linked vs code and intel R with abaqus and its completely ready for working.
U can refer to the ABAQUS user guide about the details of user defined field.
• asked a question related to Abaqus
Question
Dear all,
I want to connect the surfaces of two shell structures (shell rings). For that purpose I was thinking tie constraints. However I am not sure how to do it properly. For example how should I assign thickness offset of each shell part (bottom surface, top surface , middle surface) in order to avoid any inconsistencies?
Best regards
if you want to be spoon fed, good luck with that... already gave you ans n guidance
• asked a question related to Abaqus
Question
I model high strength concrete slab under fire using abaqus. Spalling has important effect on the fire resistance of HSC. The question is how can model the spalling the concrete in abaqus?
Mathematical modeling is possible taking into account all parameters. Of course, results can be easily obtained by programming later.
• asked a question related to Abaqus
Question
Suppose one side of the rope is closed and the other side is on the pulley. Force is applied to the free side of rope. How can I simulate this, since the rope does not bear shear force? Which element? Beam? Trus?
Hi Ghanbar,
You can use the spring for your model.
• asked a question related to Abaqus
Question
Hello
I am trying to do a sensitivity analysis by xfem Abaqus but as I am increasing the number of seeds the number of cracks increased and even decreases. which is so strange, please let me know.
Generally mesh sensitivity means the criteria when our results don't depend on mesh size. But here strange is happening, first, the number of cracks should be one but it is having more (though I understand the reason those elements simultaneously having the same initiation stress) but is this actually correct? Please let me know.
Please see the attached where I am increasing the seeds.
With regards
Biswabhanu Puhan
Please use static analysis to check mesh dependence. When a crack is terminating at interface , it is perpendicular interface crack. It shall have different singularity.
parallel interface crack has oscillating singularity.
• asked a question related to Abaqus
Question
I'm going to analysis of theorical RC beam by FE nonlinear analysis then verify the RC beam models with FEA softwares.
The best FEA software is one you know how to use. You could get good results in both of them. But, if you are starting to learn any of the software, I would suggest you to learn ANSYS.
• asked a question related to Abaqus
Question
I have a complex model in solidworks and I want to do some analysis in Abaqus. So I exported it, but everytime whenever I am doing modal analysis, I am getting the error THREE FACTORIZATIONS FAILED IN A ROW AND CONTAINS KINEMATIC COUPLING.
BUT actually I haven't applied any interaction. Also using Subspace solver, the values are zero.
Aman Kumar Can you share your Abaqus model (.inp)?
• asked a question related to Abaqus
Question
Hi all,
I am modeling a more than 100 years old truss railroad bridge using an Ansys or an Abaqus. This is a very interesting bridge to model. The members are made up of angles combined together to form a channel shape on either side of the cross-section.
Could you please advise how I could custom design this cross-section? Any advice on alternative ways of modeling such cross-section would be helpful as well.
Please see attached picture for the truss member.
Thank you
Finite element software is not powerful in terms of creating complex and custom objects. I recommend using AutoCad-3D and import the object in Abaqus.
• asked a question related to Abaqus
Question
I have taken submodel displacements from global elastic model. As the first inelastic cycle completes, I use a restart option. When i start 2nd cycle, using step=1, ABAQUS throws an error that step has to greater than step in previous run as its a restart file. If I change the step to (say 125), again ABAQUS throws an error saying it cannot find step 125 in global model. (Global model is a single cycle model with steps up to 124).
How to proceed further >?
you should refresh your textbooks regarding submodeling analysis...no software can give you response for two cycles in submodel using results from one cycle in global model... it's theoretically incorrect
• asked a question related to Abaqus
Question
Dear all,
I want to model a 3D rubber in Abaqus (dynamic explicit). When I run the model, the analysis is aborted before it starts and this error appears: “Analysis Input File Processor exited with an error”.
The Marlow model (in hyperelastic properties) is used in this modeling. It should be noted that when using other models (for example ogden model or polynomial model and etc.), this error does not appear and the analysis will be complete.
Can you please suggest a solution?
check with the warning label, it would give you some hint. Moreover, try to use some default parameters or others' set in Marlow model. Change one parameter each time. You would find the problem.
• asked a question related to Abaqus
Question
Hi, there!
I'm looking to validate an experimental uniaxial stress-strain curve of a direct tension test on a concrete cylinder using the Concrete Damaged Plasticity model in Abaqus. The material data input follows both manual and literature instructions, with a choice of stress x crack-opening displacement relation for the tension stiffening behaviour.
For a '1D' model using bar elements to represent the concrete cylinder as a "pillar", the simulation mean outputs for stress and strain are a near-perfect match to the corresponding stress-strain curve for the input data (-> this curve was derived with Cornelissen et al. exponential law together with the characteristic element length outputted by Abaqus in the message file to go from crack-opening displacements to post-peak strain values).
However, results from 2D and 3D simulations of the 'same' concrete cylinder do not match their corresponding input stress-strain curves in the least.
I have uploaded an image to better show the divergences.
Has anyone encountered something similar and would know of anything that could be the cause of this?
I appreciate any help!
Thanks!
- Make sure you perform single element tests. Don't perform small scale simulations with damage models with more elements because you won't get localized damage.
- If you already perform single element tests, this could be caused by a too small element size. If your element size is much smaller than let's say 50mm you can try to increase it to 200mm.
- Double check if your tensile behavior settings are type displacement
Check out my Matlab (>2018b) based GUI tool for easy CDP setup:
• asked a question related to Abaqus
Question
I have a plate with rectangular geometry and I want to simulate a free-free boundary condition
I try to do that with RP in the interaction module in Abaqus and I fail. I will be grateful for any help in this problem.
One trick that may work is to start the analysis with a plate constrained by elastic supports on all edges. Specify the stiffness of the elastic supports, then gradually reduce the stiffness values on all the edges. As the stiffness values approach zero, your result will approach that of the free free boundary condition.
• asked a question related to Abaqus
Question
Hello Everyone...I have been trying to reproduce the same Stress vs Strain behaviour for the Composite Material I got from a research paper to my Abaqus model. The material properties, CAD model etc., are all taken from the said research paper, but I still not making any progress. If anyone has faced and overcome this problem kindly get back to me so that I can share my isssue in detail. Thanks.
Maybe you are missing something important,I mean the value of an important parameter.Also I think a good way might be to use the charts and figure of graphs provided to extract and crosscheck some values which you are assuming as something which it is not.
• asked a question related to Abaqus
Question
I have a plate with rectangular geometry and I want to simulate a free-free boundary condition
I try to do that with RP in the interaction module in Abaqus and I fail. I will be grateful for any help with this problem.
One trick that may work is to start the analysis with a plate constrained by elastic supports on all edges. Specify the stiffness of the elastic supports, then gradually reduce the stiffness values on all the edges. As the stiffness values approach zero, your result will approach that of the free free boundary condition.
• asked a question related to Abaqus
Question
I run a 130 hrs simulation, some time I need to verify if the results are acceptable. if I could access the simulation while it is still progressing I could stop it and change the parameters if some thing goes wrong, avoiding to weight 130 hrs each time. I used to work with abaqus and I am new to Ansys
Hi Marouan,
Ansys Mechanical has quite a nice function to monitor results during the solving process of the model. When the model is solving, you can insert a "result plot tracker" which allows you to see some results (deformation, strain, stress) as the model is solving. To insert a stress plot tracker, right click on "Solution Information" in your model tree, then "insert" --> "Stress Plot Tracker". Once inserted, you can right click the Stress Plot Tracker object in your model tree and select if you want to update your results automatically or manually.
Here is a video that uses this approach with an implicit solver, but it works the same with Autodyn:
Hope that helps.
• asked a question related to Abaqus
Question
Hello everyone
I am trying to simulate a cantilever beam with direct cyclic analysis approach.
However, at step time zero, the values of stress is not zero and showing some other value.
Can anybody help me resolve the issue?
What kind of amplitude type are you using? If you want a sinusoidal form, you should configure initial amplitude and coefficient A = 0 in Edit Amplitude.
• asked a question related to Abaqus
Question
The model is very simple. The step is Dynamic Explicit. The step time is 0.01s, and the frame number is 4000(a little bit large). The odb file is only 5GB.
When I extract the data from the free body cut, it can only extract about 2000 frames and an error happens: ''Operation did not complete due to a memory allocation failure''.
My computer is the workstation. So, it's not hardware problem.
I have tried many solutions:
1. set Kernel memory limit from Tools->Options->Memory, I set the memory 1e9Mb. When outputting about 3000 frames, I met the same error.
2. set Kernel memory limit with python command: session.setValues(kernelMemoryLimit=500000), and output the results with python command. It does not work, too.
I only want to extract a very simple free body cut data from a 5Gb odb file. How to solve this problem?
Qamar Ul Islam Thank you for your answer. But I have known how to create the free body cut and report them. The problem is the error occurs when reporting a large amount of them.
• asked a question related to Abaqus
Question
How have people set up loading/boundary conditions for a quenching heat treatment model simulation? Do you apply loads or boundary conditions? Do you use an interface condition such as surface film condition and apply a film coeff (your HTC value)? I havn't been able to find any papers that really explicitly lay out how the setup is done.
The heating temperature before quenching can be set in the Load module as predefined field. for the HTC, we can use the surface film condition-f(T) in the interaction module。
• asked a question related to Abaqus
Question
hello
I'm trying to modeling composite panel and damage tolerance analysis but I don't know how to set step and what I should achieve at the end of work?
The below-mentioned literature may be useful to you.
• asked a question related to Abaqus
Question
Hello everyone,
I am trying to model the damage evolution in a bolted connection in Abaqus with element deletion option. I have the stress-strain data for the bolt but not sure how to extract the required inputs necessary to model the damage or the fracture in the bolt. Below is a short explination for how I did the modeling:
Material - mechanical - damage for ductile metals - shear damage.
Then, a window opened asking to enter the material parameter (ks), fracture strain, shear stress ratio, and strain rate. (see attached photo).
After that in order to model the damage evolution, I selected sup-option and it asked me for the displacement at failure. (see the attached photo).
I do not know how or from where I can get these values. I tried to read about the damage mechanics but could not get how obtain these values. Are there any resources that can help or any examples or explanations?
Thank you,
Maybe you can find the definition and example here:
Bests
• asked a question related to Abaqus
Question
I want to perform a nonlinear dynamic analysis (DYNAMIC IMPLICIT step) of a RC column under earthquake excitation in Abaqus (Standard or CAE).
While there is no problem with my simulation using MATERIAL DAMPING options with default materials (like Concrete Damaged Plasticity), I can not apply Rayleigh damping when I use an User Material for Concrete (see screenshot). The following error occurs:
***ERROR: A MATERIAL PROPERTY THAT IS NOT ALLOWED WITH A USER MATERIAL HAS BEEN SPECIFIED
So apparently it is simply not possible to use it with a UMAT.
Are you aware of other possibilities to use Rayleigh damping (stiffness & mass proportional damping) with a UMAT?
I have read about GLOBAL DAMPING, but this is not applicable for direct solution procedures like DYNAMIC IMPLICIT.
What properties of a Rayleigh damping do you want to apply?
• asked a question related to Abaqus
Question
I want to create a beam(3d wire) in abaqus with a conical cross-section and different thickness. However in cross-section modulus there is no something like that. How could I do that? Please give as much information as possible
For complex objects with variable cross-sections, I recommend sketching the object in AutoCAD then import the object into the Abaqus.
this is the easiest way for all kinds of complex shapes.
• asked a question related to Abaqus
Question
I'm trying to use the Conventional Shell for the I-profile to validate a model in a paper. However, there is no value (i.e, number of plies, thickness, orientation) been provided. So, is there another method to get these inputs?
Thanks for updating me, and I really appreciate your cooperation.
Best regards,
Aziz
• asked a question related to Abaqus
Question
Dear all,
I want to connect two parts in abaqus. One is a cylindrical shell and the other is a solid deformable ring. The ring will act as a stiffener for the cylindrical shell. My question is what is the appropriate way to connect them?
1) Should I connect the parts in Assembly modulus and then merge them together?
2) Should I design both in Part modulus?
3)Or should I tie them with a solid-to-shell constraint?
What is the best strategy? How the results will be affected by choosing one of these 3 methods?
• asked a question related to Abaqus
Question
I try to compare a neo-hookean material implemented in Abaqus with one made by a user subroutine (UMAT) in fortran.
The user subroutine came from the same author as the article
Isotropic hyperelasticity in principal stretches: explicit elasticity tensors and numerical implementation
My material is neo-hookean, incompressible, with material property C1 = 0.005
The problem is that when I try to pull my beam, with abaqus neo-hookean simulation, my beam becomes thinner (what is expected), but when I try to do it with my UMAT, the beam becomes thicker (not expected)
You can see my results in the image "Comparison umat and abaqus.JPG"
Do you know why I have this incoherence ?
Thanks
As you may already know, the only important thing in your UMAT is the definition of the Cauchy stress tensor. Your final results only depend on this tensor. The tangent modulus (DDSDDE) only affects the convergence of your model. So, if your stress tensor is defined rightly, you can expect to have correct answers, regardless of the accuracy of DDSDDE.
You should check the definition of the Cauchy stress in your UMAT, and since you're using the simple neo-Hookean model, this should not be a hard thing to do. I suspect that you may have mistaken a sign (+/-) in your Cauchy stress definition.
• asked a question related to Abaqus
Question
1. I am trying to simulate a pure shear loading on a cubical block with a single brick element.
2. While I am simulating the other stress components are not getting zeros.
3. Can anyone suggest to me how to apply boundary conditions to simulate a pure shear loading on a 3-D cubical block?
In order to correctly apply a boundary condition of simple/pure shear a possible solution is to impose the so-called Periodic Boundary Condition (PBC). You can find an analytical formulation of this boundary condition in the following paper:
Micromechanical analysis of polymer composites reinforced by unidirectional fibres: Part II – Micromechanical analyses by A.R. Melro,  P.P. Camanho, F.M. Andrade Pires, S.T. Pinho.
or in this book:
Finite Element Analysis of Composite Materials using Abaqus, by E.J. Barbero.
For a correct numerical implementation you should create a set for each edge, vertices and face of your brick, and for each of them extract a single set for each node contained. After that, you should create an equation for each corresponding node on the opposite edges/faces.
I think that the knowledge of python scripting for the implementation of these conditions can be helpful.
Note: This is the genel approach to get a simple/pure shear BC, but in your case it will be easier since you do not have nodes on the faces.
• asked a question related to Abaqus
Question
Hi,
I'm trying to simulate an orthogonal cutting operation on Abaqus, I'm using Dynamic explicit analysis. I have defined my tool as Rigid analytical body and the workpiece is of Ti6Al4V, I have not assigned any material to the tool.
My problem is that the simulation runs untill 50 increments and completes, but the cutting operation dosen't take place, I have attached images for reference. I have used 0.025 as time period and 0.0005 as time increment. Any variation of time increment results in an error.
I agree with Debijyoti.
I would also recommend to simplify as much as possible.
Get rid of the adaptive meshing at first until things run reasonably well.
Let the tool move much further to see whether your material will actually fail as desired.
If not, create a simple model (e.g. a tensile specimen) with just one or a small number of elements and check on this whether your failure model works as expected.
If you are restricted to the student version, perhaps only model the chip; create a small, thin layer of failure elements at the bottom of the model that are to be cut by the tool and fix their lower nodes with a bc. That will allow you to have more elements in the chip where deformation is strong. (I assume that you are doing this only for the purpose of understanding - there is no way to create a realistic machining model with 1000 nodes.)
• asked a question related to Abaqus
Question
I am writing Abaqus USDFLD to change my Young's Modulus after strain crosses the value of 0.035 and start deleting elements when strain value crosses 0.05 in a heterogeneous model (composed of two different sections).
I am using linear degradation; (1-D)*E where D goes from 0 to 1 as strain goes from 0.035 to 0.05.
See the image below, and I want that degradation to follow the linear curve.
So, my problem is how do I infer the young's modulus (E) and which section I am at when I use the GETVRM subroutine to get values at the integration point? I have checked the Abaqus documentation but couldn't find the method to get E at the integration points. Is there any alternative way to accomplish this?
And I am assuming we need to supply the same subroutine file (.for) for all the material, which gives rise to this issue, and I am unsure how to supply different subroutine files for all the different sections we have. Is this possible?
Hi Ayush,
I suggest you add a Y constraint over the whole of the bottom of the cube instead of just the edges. I also suggest you add two X constraints at different points on the Z axis and one Z constraint somewhere on the X axis.
And please make an exception and call me Simon even though I am over the hill.
Regards,
Simon
• asked a question related to Abaqus
Question
Hi all,
I have modelled a high-speed train running on a soil medium. The waves hitting the boundary and reflecting back. I used infinite elements but they somehow cause a residual displacement and viscous elements elements seem to reflect some waves back. Do you have any other recommendations how to implement absorbing/non-reflecting boundary?
Regards, Ahmet
#abaqus
Hi Ahmet
Code Aster/MISS 3D should handle this case for you and is state-of-the-art.
That the combo execute inside Python does not require any coding, it just makes life simpler if you want to tweak anything.
For Rayleigh Waves, I believe that infinite elements are highly reflecting.
If you want to continue using Abaqus, perhaps you may want to use the approach used herein?
Sincerely
Claes
• asked a question related to Abaqus
Question
Hi,
I am simulating rock cutting process in Abaqus. The simulation starts with a brick shaped rock (flat top surface) and one of the responses I am interested in is the removed volume of rock at the end of the cutting process. I should note that there is a damage model defined and the elements get deleted accordingly.
Using output variable STATUS does not work unfortunately, since some elements simply fly away without being deleted (when I checked the STATUS value for those flying elements, I can see they are set to 1). In other words, some rock particles simply get removed without being deleted.
I tried using a displacement criteria, essentially checking displacement of nodes and if they are above a threshold, I consider them as "removed". However, I was wondering if there would be a more elegant and robust way to do this. Is there for example another output variable in Abaqus, which would tell me element IDs for those elements which lost connectivity to the rest of the model.
Best regards,
Murat
why you not divide your model to step , and one of them will be the excavation , and by using the interactions , make a deactivated for this step .
• asked a question related to Abaqus
Question
Hello! can someone help me please?
I'm simulating a tensile test with flat specimen in Abaqus explicit using JC plasticity and JC failure criterion. I need the strainxstress curve, a need the total strain not only the plastic strain, but in the "creat xy data --> ODB field output" doesn't appear the E. But I have selected it in the "Field Output Request"
Hi Juliana, you should settle down Logarithmic Strain (LE) in field output and consider (LE) results after completed running
• asked a question related to Abaqus
Question
The monitor shows I have calculated 29 steps, but I can only see 7 steps from the results though the visualization. I use the abaqus 2018 explicit solver.
Thank you
If this does not work, you can request an output for each step.
• asked a question related to Abaqus
Question
Vfric subroutine offers to prescribe a model based on slip rate and contact pressure. However I want to include strain and use my own model for contact pressure and slip rate. Is there anybody who has experience in it?
Vfric is the subroutine to change friction values but you cannot calculate based on the contact pressure.
• asked a question related to Abaqus
Question
Dear all,
How to find the cooling rate in welding using Abaqus Subroutine for incorporating the fraction of Martensite due to phase transformation.
Hi Anoop,
Simon
• asked a question related to Abaqus
Question
Hello All,
i have the following warning message displayed for an explicit simulation. What could be the reason for this warning?
I am using amplitude to define the displacement of "Welle" in step 2 and then the "Welle" is not moving in step 3.
***WARNING: The option *BOUNDARY,TYPE=DISPLACEMENT has been used with a jump
of 1.80000 at DOF 2 of node WELLE-1.1 at the beginning of step 3.
***NOTE: Displacement boundary conditions with prescribed jumps will be
enforced as defined, minus the value of the jump. The jumps might be
caused by:
1) A discontinuous *BOUNDARY, TYPE=DISPLACEMENT definition.
2) Initial contact adjustments at the beginning of the first
step.
3) Locating the rigid body reference node with the POSITION=CENTER
OF MASS parameter.
4) Use of an amplitude curve to prescribe the boundary conditions
in an analysis with a large number of increments; precision
error may cause a small drift from the prescribed displacement.
Did you find the solution?
• asked a question related to Abaqus
Question
I am rather new to Abaqus/Isight. So, what I did was to create a model in Abaqus (using the GUI), and then try to optimize the geometry of one of my parts in Isight using the .cae file as input.
However, Isight doesnt run Abaqus unless I use the .inp file as input. Thus, I was wondering if there is a way to access the dimensions of my pieces and name them using the .inp file generated by abaqus.
Help is greatly appreciated !! :)
Pay attention to such Abaqus capability as parameterization: https://abaqus-docs.mit.edu/2017/English/SIMACAEMODRefMap/simamod-c-parinput.htm
• asked a question related to Abaqus
Question
Dear all Abaqus users,
I want to connect a cylindrical shell with a solid ring, where the ring simulates the flange. For this purpose I was thinking to the shell-to-solid constraint. However, some qiestions arise and I would like to ask for your help.
I am a little bit confused on how to connect solid and shell and how shell's offset affect the connection.
In figure 1 ,below there is a sketch of the model.
For the connection should I use the green area which it's width(0.016 m) is the same as the cylinrical shell's thickness or should I use the whole flange's surface?
Any help will be appreciated.
Quvonchbek Rakhimberdiyev Sir, that is a sketch of my abaqus model. However, with all the respect, you didn't answer to my question
• asked a question related to Abaqus
Question
I asked this a while ago, but decided to dig a little deeper before asking again. When using the Abaqus JH1/JH2/JHB Johnson-Holquist models for brittle materials, the FS (Failure strain) element removal threshold value is extremely important to ensure the simulation matches experiments, yet in practically all of the studies I have looked into, none of them have reported this FS value despite its major significance.
In the Abaqus documentation, FS was given as 0.2 for SiC. Subsequent studies on alumina have claimed that this FS value needed to be adjusted for accuracy, but then never reported it (e.g. Liu et al. 2015; Khan et al., 2021; Guo et al. 2020; Goda & Girardot 2021).
If anyone can shed some light about what this FS value is meant to be and how to obtain it experimentally or analytically, I would love to understand it better. Thank you!
Hi Zherui,
FS is a parameter that provides flexibility on correlating numerical with experimental results. I reproduced the formulation of plastic strain in the attached file, as one can see the FS value depends on the values of D_1 and D_2 and the negative hydrostatic pressure strength, T.
If one is confident with the remaining parameter values, then an acceptable strategy for deciding on FS value would be to minimise the error between experimental and numerical results based on multiple references. In the case of impact, the ballistic limit, projectile velocity-time profile and residual projectile velocity for instance.
Please consider [1] for a comprehensive analysis of the effects of FS.
For HDCs, FS=0.6 was previously used yielding satisfactory results for the ballistic limit and residual velocity and FS=1.5 for DoP.
In addition, FS=3.0 was used in [2] to validate Johnson and Holmquist [3] cases,
successfully reproducing the JH results.
Regards,
Theo
2. : Implementation and Validation of the Johnson-Holmquist Ceramic Material Model in LS-Dyna
3. JOHNSON, G.R. AND HOLMQUIST, T.J., (1994) “An improved computational constitutive model for brittle materials”, High-Pressure Science and Technology – 1994, American Institute of Physics
• asked a question related to Abaqus
Question
For the case of crack propagation abaqus provides both the above methods. What is the difference between the two?
The usual answers seen are just mentioning that one uses CZM and another LEFM.
If anyone could give more details into them it would be really helpful. Since it is not yet
available in literature if both offer the same accuracy or not.
I am not sure but i think LEFM (Linear elastic fracture mechanics) model does not account for resistance caused by blunting of crack tip.
• asked a question related to Abaqus
Question
I am trying to import a material library (.lib) in Abaqus/CAE 2017 but i don´t know how can i do it (I only know how to create a material library).
Would anyone know how to do it?
You need a plugin for that check in abaqus plug in menu, if it is already available other wise download and install it.
• asked a question related to Abaqus
Question
Hi every one
I am working on a model with stretching and bending in Abaqus. In my first step I apply the stretching to my model and in the second step I apply the bending. The magnitude of kinetic energy is small in 1st step but it increases significantly in bending step and exceeds the internal energy. Is there anyway to reduce the magnitude of kinetic energy? Also I should say that I have not used the mass scaling!
Thanks
Good luck
• asked a question related to Abaqus
Question
I am trying to do an UMAT subroutine, but when i evaluated it in Abaqus Command many errors were appeared.
Would anyone know how to solve them?
I attache an image of the errors
this is due to some un defined variables in the code. try to solve each error one by one. check .sta file for for information.
• asked a question related to Abaqus
Question
I am trying to carry out a study in which in the first step the cylindrical part only moves in the Y axis (for which I apply a BC limiting U1, U3, UR1, UR2 and UR3 = 0 and a pressure on the upper face of that part which cause its movement), however when in the next step I want the part to stay in the position in which it ended in the previous step (that is, it does not return to its initial position) I cannot do it.
I have tried to limit the 6 degrees of freedom in that 2nd step with a BC, but that causes that even in the 1st step there is no movement.
Would someone know how to do it?
I attach images and a video for a better understanding