LS-DYNA TUTORIAL 18: Sphere Drop on Water with ALE method

[@tanishashrestha7053]

Grateful to you again for posting such amazing educational videos!! Indeed a powerful explanation! Keep posting and enlightening us!

[@AmeenTopa]

You are most welcome. Please share my videos if you find them useful. 😬

[@jiayueding4314]

Hi Ameen, i watched your tutorials a lot recently since i am doing my master thesis about FSI problems using ALE, ICFD and CESE solvers in LS-Dyna. Thanks a lot and keep going!

[@AmeenTopa]

Best of luck! I have not posted any tutorials on ICFD or CESE yet. But if you need personal consultation, I am available. But that will cost you at least one journal paper publication. XD

[@jiayueding4314]

@@AmeenTopa Thanks! Honestly i really need some helps since i am a newbie in LS-Dyna. But i am trying my best to overcome the difficulties and thanks to your tutorials it helps me a lot! I will message you again if i am really into "troubles" :)

[@AmeenTopa]

@@jiayueding4314 Yeah sure. We can communicate through email. Sometimes, I don't get the notification here on KZbin.

[@ahdshsah4869]

Hi Ameen, Thanks a lot for the video, It helps a lot in my current project, great appreciated. One tiny error may need to be corrected, the dynamic viscosity of water(mu) should be 1.0e-9 Gpa.ms not 1.0e-8

[@AmeenTopa]

Yes you are correct. I must have mistakenly converted from 10^−3 Pa*s. Thanks for pointing it out!

[@ahdshsah4869]

​@@AmeenTopa Yes, I am currently moving from Ansys to Dyna to simulate some fluid slosh problem and the small particle movement inside the tank via the SPH method, Dyna is much more difficult to use compared to Ansys and I keeping make mistake when I convert the unit:(

[@AmeenTopa]

@@ahdshsah4869 You are right. LS-DYNA is not as user friendly as ANSYS and not many materials/tutorials are available online.

[@lalinlamCambodia]

@@AmeenTopa really? I haven't checked ANSYS yet, but for now I think LS-DYNA was already user friendly :D. or maybe you will have some videos with ANSYS so that we can compare :D

[@AmeenTopa]

@@lalinlamCambodia I don't plan to upload any ANSYS videos. There are many already out there. Check out this example: he did a similar thing (ball impact on water, but at angle), in less than 5 mins! :D

[@sumankumarrr]

I am grateful to you for this video. Respect for you! Excellent

[@AmeenTopa]

Please share my videos so more people can learn ☺️

[@ajaygoyal2411]

Amazing, Sir Ameen. Very informative. Thanks for sharing. It could be great if you may make some videos regarding mesh convergence, which is vital in ALE and FSI problems. I have noticed severe fluctuations in the results if mesh size changes from 10 to 5 or to 1. Some literature suggests that mesh size is 1 by 1 by 1 is good for non-linear FSI problems but it severely increases the computation time (2-3 days). The memory required to execute such problems is nearly 250GB. This poses limitations for use of ALE method but the results are promising.So, it will be great if you can share your vision regarding mesh size and mesh convergence for FSI related problems. Thanks and good luck.

[@AmeenTopa]

Thank you for your input. To perform mesh convergence analysis, the same steps in this video can be followed. The only thing is that you make the elements smaller at the meshing step. I used 10x10x10, one can use then 8x8x8, then 6x6x6 etc. Then the rest of the steps are the same. There is no one best solution for all cases. Optimum mesh size depends on the size of the domain. If the domain is very small only 10x10x10, then size 1x1x1 size will be very coarse mesh.

[@minli-u5k]

Hello, may I ask how to extract the impact force of cylindrical shell after axial impact in LS-PREPOST? Can I multiply the overall acceleration times the mass to find the impact force? Thank you very much for your answer!

[@AmeenTopa]

If the impact is done using: 1- contact with a plate, then use database rcforc 2- contact with a *RIGIDWALL, then use database rcforc 3- prescribed motion node set, then use database bndout Your question lacks details about your model, the above is all that I can recommend. Watch my tutorials: 1- Axial Tube Crush 2- Post processing with LS-Prepost 3- How to plot f-d diagram. I hope this helps.

[@minli-u5k]

@@AmeenTopa Thank you very much, I will go to watch the video learning you recommended

[@dequan-g2z]

Very excellent tutorial! May I ask have you tried using Structured ALE method to conduct FSI simulation?

[@AmeenTopa]

@@dequan-g2z thank you. Yes, I use it as well. But I don't have any tutorials on it in my channel .

[@dequan-g2z]

@@AmeenTopa Thank you for the reply!May I ask u, how to create a cylindrical SALE area? Thank you again!

[@AmeenTopa]

@@dequan-g2z you need to make a box then trim the SALE domain.

[@dequan-g2z]

@@AmeenTopa Is it conducted by *ALE_STRUCTURED_MESH_TRIM?

[@AmeenTopa]

@@dequan-g2z bingo! That's the one.

[@МаксимГуляев-з7щ]

Thank you from Russia!

[@AmeenTopa]

You are welcome from Malaysia!!

[@abrahammorenoreyna2758]

Amazing!! Thanks for you job!!

[@AmeenTopa]

Thanks! Please share my videos so more can benefit from it :)

[@こぱんだ-k4o]

In the video it was a vacuum, but if you want to recreate the general atmospheric conditions, is it better to use MAT_NULL instead of MAT_VACUUM?

[@AmeenTopa]

Yes. And you need EOS for that. Check my tutorial of Blast on Concrete Beam

[@CuongPeugoet]

thank you very much! it's very helpful! How you can remesh? Should we start from the beginning?

[@AmeenTopa]

If you want to refine the mesh, you can go to Element Tools, Elements Edit, Split. If you want to make the mesh coarser, then you have to delete and create a new mesh.

[@CuongPeugoet]

@@AmeenTopa thank you very much! it works

[@AmeenTopa]

@CuongPeugoet you are welcome.

[@lenincruz299]

Hello, when I run the analysis it appears *** Note history variables in lsprepost --> Fcomp -->misc history variable 1 to 3 belongs to AMMG 1 history variable 4 to 13 belongs to AMMG 2 what could be the error?

[@AmeenTopa]

If this does not causes a termination error, then you can simply ignore it.

[@miguelvega1830]

Thanks for all your help with every video you make, dear Ameen. I've tried to do every step you did in the tutorial but I just get the sphere going into the water but there's no any movement in the water. Please, what might be doing wrong?

[@AmeenTopa]

If you defined CONSTRAINED LAGRANGE IN SOLID correctly, then the interaction between the ball and the water should work. It could also that your model is perfectly fine, you only need to change the display setting to see the ALE (fluid) parts.

[@miguelvega1830]

@@AmeenTopa Thank you very much for answering so soon. The input in the definition of CONSTRAINED LAGRANGE IN SOLID have different names in my LS-Prepost, I don't know if it is related to the problem. Instead of SLAVE, I have LSTRSID. Instead of MASTER, I have ALESID. Instead of SSTYP, I have LSTRSTYP. And instead of MSTYP, I have ALESTYP. The rest is the same. I'm not sure if I need to change the display setting, because I do see the water but it doesn't move at all. Or maybe I'm not considering a setting that could help. I will thank you any comment you can give me, dear Ameen.

[@ramkrishna4033]

In typical FSI problems, the solid part needs to be subtracted from the background fluid domain using a boolean operation. That way, there is a well defined fluid solid interface that is formed. That was not done in this FSI tutorial with LS DYNA. So, how does the coupling work here?

[@AmeenTopa]

Coupling between solid and fluid is done using Constrained Lagrange in Solid. The approach used in this video is called Arbitration Lagrange Eulerian. This is not like a typical CFD method, so the overlap between the solid and the fluid is allowed.

[@pritamghosh59]

I have followed the same steps, but I am getting dbfsi as zero. Any possible solution

[@AmeenTopa]

Please recheck again your steps. Make sure to 'insert' before you click on 'Accept'

[@pritamghosh59]

@@AmeenTopa thank you for your response, would you mind if I mail you the k file.

[@lalinlamCambodia]

Got the same problem here. Have you solve it yet?

[@lalinlamCambodia]

@@AmeenTopa i got the same problem as the guy above. I already click insert when I add the ALE part set to the database FSI.

[@AmeenTopa]

@@lalinlamCambodia You follow the steps exactly? You click on insert but nothing changes?

[@ddnull9681]

Why does the calculation fail (appearance of a negative volume mesh) after adding a no-reflection boundary

[@AmeenTopa]

I am not sure. I have did slamming simulation with nonreflective boundaries and didn't get negative volume error. So, the cause might be something else or it could be improper definition of the non reflecting boundaries.

[@ddnull9681]

@@AmeenTopaWhich surfaces did you set as non-reflective surfaces? Except for the two symmetry planes, I set the remaining four to be non-reflective, and the calculation would be interrupted by the negative volume mesh.

[@AmeenTopa]

@@ddnull9681 Yes, it should be only the segments of the four outer faces. Have you tried the simulation without the non-reflective boundaries?

[@ddnull9681]

@@AmeenTopa I have tried not setting the no-reflection boundary condition and it was very successful, but I would like to try the no-reflection boundary condition, and I read the paper to set this condition.

[@AmeenTopa]

Okay. It it worked without the nonreflective boundaries, then it means only one thing. There is a mistake in the way that you defined this boundary conditions.

[@raviprasanth3675]

Can u please do analysis ALE method with volume fraction geometry technique

[@vaibhavnmete]

The video was amazing. Sir kindly provide how to run .k file or runner for ls dyna or any third party runner.

[@AmeenTopa]

Hi there, to run LS-DYNA files, you need to have the LS-DYNA manager. You can contact your local distributor to get the software.

[@shout_gopchang]

hello how can I move ALE parts? I'm doing bird strike on aircraft leading edge, but my ALE bird is not moving. I used INITIAL_VELOCITY and used set_node. My bird is not moving at all.. help me

[@AmeenTopa]

You need ALE reference node set. I have did this in my second ALE tutorial (sloshing), you can check that out!

[@shizhou1106]

hi，Thanks for sharing.I have a problem about UNIT SYSTEM, what is the unit of water and air in EOS and MAT NULL keyword?

[@AmeenTopa]

In this video, I used kg-mm-ms unit system.

[@shiqizhang3532]

it's wonderful! I'm just looking for a quarter model to simulate video teaching And s-ale grid division great!!!!😀

[@AmeenTopa]

Unfortunately, I don't have any videos on S-ALE yet. But if you understand the idea of how to make a quarter model, that would be enough.

[@shiqizhang3532]

@@AmeenTopa it's enough i'm so happy to get this skill :)

[@AmeenTopa]

Please share my channel so more people can learn! :)

[@shiqizhang3532]

@@AmeenTopa it's ok

[@dakshitha92]

🙋🏻‍♂️🥂🥂🥂👏🏻 thank you for sharing

[@AmeenTopa]

No problem. And thanks for sharing the video :)

[@sahilll7694]

Sir do we need to know time step calculations and everything elese behind all these stimulations

[@AmeenTopa]

No you don't. But it is highly recommended that you do.

[@marcinsmulski7632]

why the MU Value is 1e-8? The value in 20 Celsius degree its about 1004*10^-6 Pa*s which means it should be 1.004*10^-3 Pa*s . The linked article also has an mistake. They write about kinematic viscosity and the uses the unis Pa*s.. Can anybody explain it to me?:)

[@AmeenTopa]

The article is correct. I converted the value incorrectly. :D

[@DariaAndrievskaia]

Thanks for this amazing and very helpful tutorial! For my project, I simulate an impact of the projectile on the moon surface. So, I suppose I can use this model as an example but change material and its properties for the ground (instead of water)? I want to have similar results since I should analyze the ejecta movement (particles after the collision). Thank you in advance!

[@AmeenTopa]

You are welcome. Please share my videos if you find them useful. Yes it is possible. That's a very interesting topic you are doing.

[@DariaAndrievskaia]

@@AmeenTopa Thanks! Actually, during running, I've got a problem: the sphere just penetrates the water without any interaction (just go through as some kind of ghost :( ) Do you know what could be a problem here?

[@AmeenTopa]

Are you following the tutorial exactly? If yes, make sure that you are showing the ALE part when post processing, not the solid part.

[@DariaAndrievskaia]

@@AmeenTopa Yes, step by step. I'm showing the ALE part for the water and the solid one for the impactor. I've checked, everything's the same but the problem still stands. I will try to redo it from scratch.

[@DariaAndrievskaia]

​@@AmeenTopa Good afternoon! I tried to redo the whole tutorial from scratch but unfortunately, I see the same result.. Could you please help me out?

[@ajaygoyal2411]

Moreover, can we edit the graphs directly in LS Dyna? Suppose, someone wishes to (a) edit the title of axes, (b) the color of the plot, (c) the size of the graph, (d) font and style of the text, (e) text insertion. My usual practice is to import the generated data (say acceleration vs time of ball) from the generated file to excel, then remove the unnecessary data, plot graph in excel, or MATLAB with font, style, size of my interest. But it is a tedious task if I need to plot numerous graphs. So is there is a way to edit the graph in LS Dyna which can be done using excel or MATLAB?

[@AmeenTopa]

Yes. We have some options for (a) - (d), regarding the text insertion, I am not sure. I think I have covered few of these options in my tutorial "tensile test with solid elements"

[@TheJustsoso]

Hi，Ameen，thank you a lot！I followed your video and the simulation is done，but the ALE deformation results in LS-PrePost can't be shown，when I click F2 , the Fluid(ale) button can't be checked，is there any way to find the problem？

[@TheJustsoso]

it point out ”No valid volume fractions for elements in model“

[@AmeenTopa]

Have you defined ALE Multi Material Group?

[@raviprasanth3675]

Initial detonation also with that

[@AmeenTopa]

Did I use Initial Detonation in this video?

[@raviprasanth3675]

No sir

[@sahilll7694]

Sir can we do some chemical reactions ex lithum in water? Result is quite explosive

[@AmeenTopa]

Yes. There is the CHEMISTRY solver recently available in LS-DYNA. I never used it before but it can do reaction simulation.

[@leungeric1803]

Thanks Sir, may I ask why we need to add vacuum or air layer above the water surface? What happen if we only have water layer with ALE?

[@AmeenTopa]

Vacuum or air is the domain in which the water will move through. If it is not there, then the splashing effect will not happen. Check out the slides that attached in the description, it will help you understand the concept of ALE. :)

[@leungeric1803]

@@AmeenTopa thanks sir!

[@AmeenTopa]

No problem. Please share my videos so it can reach more people

[@mohankrishna1854]

If you have time can you make a video on control card we use in ls-dyns

[@AmeenTopa]

First, that's a very general question. We have lots of keywords under control cards. I have already covered the most commonly used ones. So check out the rest of my channel. Second, unfortunately I don't have time to make any requests at the moment. For personal help, we can communicate through email. But that is possible only for research-based projects.

[@mohankrishna1854]

Ok thank you bro

[@lirilsilvi]

Can it apply contact angle to the sphere's surface and see its effect on splash dynamics?

[@AmeenTopa]

I don't understand what do you mean by 'contact angle'. I request elaboration.

[@lirilsilvi]

@@AmeenTopaIt means the splash created due to impact of sphere depends on the surface's contact angle. More clear explanation is given here: www.nature.com/articles/nphys545 Can it implement contact angle of surface with water and air in this solver ?

[@AmeenTopa]

@@lirilsilvi Now I saw the paper, I understand what you mean. We drop the sphere at different scenarios, the outcome will be different splashing and different contact angles. Hence, the contact angle is the output. Therefore, your first comment was confusing. We do not apply contact angle. That's not our input. We input the water properties, sphere properties, drop velocity etc and the OUTCOME is the contact angle. Yes, we can get matching results with the experiment in the paper you referenced in your previous comment. However, we need to use finer mesh (more elements are needed for the simulation to be able to capture the splashing details). And it might be more feasible to use the ICFD solver due to its capability of adaptive mesh refinement.

[@pranavoak1508]

How do we introduce buoyancy effects in the simulation??

[@AmeenTopa]

Buoyancy is an effect due to gravity. So if gravity is defined, then you should be able to see buoyancy in your model.

[@hemanthputha4792]

Sir, How to define the gravity in to the problem ?

[@minli-u5k]

Hello, I am doing collisions in LS-DYNA with different water depths, but it seems that setting different water depths has no effect on the collision results, I am very distressed, I will leave my QQ email here, when you see it, I hope to have the opportunity to discuss with you

[@AmeenTopa]

Hello there, Make sure to set the boundaries of the water correctly. If you did not define the boundary conditions for the water, or you make them non reflective boundaries, there will be no difference in the results. For personal consultation, there will be some fees. If that's fine with you, please contact my team at [e-mail]

[@minli-u5k]

@@AmeenTopa Thank you very much, I did set up infinite water, but what I want to simulate is the collision in different water depths in the deep sea, so not only do I need infinite water, but I also need to get different collision results, which is a little difficult for me

[@AmeenTopa]

@@minli-u5k if the water is infinite, then the size of the water domain won't make a difference in the results..

[@minli-u5k]

@@AmeenTopa Thank you very much for your patience to answer my question, I am using ALE_AMBIENT_HYDROSTATIC and INITIAL_ALE_HYDROSTATIC these two keywords to set the hydrostatic pressure, set different Pbase to achieve different water depth pressure size, I now go to delete NONE_REFLECTING and then see what the results will be。

[@minli-u5k]

@@AmeenTopa Hi, yesterday I removed the BOUNDARY_NONE_REFLECTION and calculated the results for two different hydrostatic pressures, the final results are still exactly the same, I am now a bit lost, I am a first year graduate student, I just learned LS-DYNA software, there are still a lot of things I do not understand, your videos really helped me a lot.

[@蔡青芳-l2w]

hi,thanks your video I do this part and add a gravity but my water ALE part will flow away, like the ALE part no boundary, l want to know if have some way to solve it

[@AmeenTopa]

Hi, in CONTROL ALE, try EBC = 1 or 2. Give this a try and let me know if it works!

[@蔡青芳-l2w]

@@AmeenTopa it works!! thanks

[@AmeenTopa]

@@蔡青芳-l2w You can upload the video in KZbin. Or, you can put your email here and I will follow up in my free time

[@蔡青芳-l2w]

@@AmeenTopa [e-mail]

[@AmeenTopa]

Noted. Please check your inbox

[@jiayueding4314]

hey Ameen, how can I make the velocity of this rigid sphere always constant during the falling process, for example i want to make the velocity always 2m/s?

[@AmeenTopa]

Hi Felix, you use BOUNDARY PRESCIRBED MOTIION RIGID. I have used this keyword in few of my other tutorials like TENSILE TEST tutorial or TORSION tutorial.

[@jiayueding4314]

@@AmeenTopa what if the material of this sphere is elastic instead of rigid body?

[@AmeenTopa]

@@jiayueding4314 If the ball if elastic, then you need to use Prescribed Motion Set and you define a node set on which the velocity is applied. These node set will have the constant velocity while the rest of the ball will have slightly less than 2m/s due to the deformation. Another way, is to use Prescribed Motion Node and apply the velocity on the node which lies on the sphere center of mass. Here, the entire ball will be deformable, but it might cause instabilities if the ball is very soft or the velocity is very high.

[@jiayueding4314]

@@AmeenTopa thank you so much for the information! I am just confused why lsprepost dont have an option for simply applying a constant velocity on the part.

[@AmeenTopa]

@@jiayueding4314 You need to define the velocity on a curve. If your curve is horizontal and straight, then your velocity is constant. If it is inclined at a slope, then the velocity is increasing at a constant rate. If the curve is a sine curve, then the velocity goes up and down. So you see, like this you will have more options on how to apply the velocity. You say "an option for simply applying constant velocity". If that's option is there, then you cannot have a variable velocity as explained above.

[@andrea2128]

great Ameen!!!!!!!!!

[@AmeenTopa]

Please share my videos so more can benefit! :)

[@Amir-hg1yh]

Hi. Could you record a tutorial about 2d axisymmetric simulation of a bullet penetration in ls dyna? I have found some articles,you could message me to share them with you on gmail.

[@AmeenTopa]

I have done a simulation on axisymmetric metal forming. You can check that one to know what keyword to use for axisymmetric simulations.

[@Amir-hg1yh]

@@AmeenTopa thanks much. im working on it but i dont know what are some of the keywords you have used.it would be good if you do a favor and release a tutorial about 2d and axisymmetric simulations and tell us about the keywords and the reason of their use.thanks again,your guidance was very useful