Tutorial of DEM Simulation using LIGGGHTS: Complete Workflow of a silo particle simulation
Вставка
- Опубліковано 12 лип 2024
- www.engineerdo.com/trainings/...
This Video shows the complete workflow of a Discrete Element Method (DEM) Simulation using the open source software LIGGGHTS. The tutorial shows a silo simulation including the filling and discharge processes.
The following parts are contained:
Work Flow and desired programs: 0:51
CAD Modeling and STL Export: 1:35
LIGGGHTS Script Explanation: 7:30
How to run LIGGGHTS: 18:20
Post Processing by Paraview: 18:44
Animation: 20:21
Link to "Compiling LIGGGHTS (DEM) 3.6.0 on Fedora": • Compiling LIGGGHTS (DE... - Наука та технологія
Guys, you are the best! The tutorial is amazing, Congratulations!
Hey Eduardo, thanks for your amazing feedback!
Thanks for the tutorial! A video explaining about time step will be extremely helpful.
Hey Ronaldo, Glad it was helpful! What do you want to know about the time step?
Cheers, EngineerDo
@@engineerdo6568 A basic guide would be ok. How to choose the proper time step to have good stability and not occupy a lot of memory. What parameters influence the value of the time step. Is the time step the same for different contact models?. What is the Rayleigh time step?.
Very good tutorial! Hope to see more :)
Excellent tutorial.
Thanks for your feedback
alhamdulillah dapet tutorial tepat
This was great!
Hey Ence444, thank you!
@@engineerdo6568 Hey for the post processing tip, I believe you need the dump to vtk reader plugin. Viewers could change the particle dump portion of the script to "dump dmpparticle all custom/vtk ${dumpstep} post/particles_*.vtk id type x y z vx vy vz fx fy fz radius mass" to output the vtk format.
@ENCE 444 Hey Ence444, thanks for sharing your feedback. We appreciate.This would be also a good option to use the vtk format. We try to mention it.
Otherwise the plugin reader can be download and install via:
www.cfdem.com/paraview-plugin
Thank you :)
Amazing tutorial🙏🏻please make a tutorial on coupling cfd dem, like particles also interact with buoyancy, drag forces.
Thanks Amritpal, the tutorial you are looking for, can be found here ua-cam.com/video/HsTRkC5EIOw/v-deo.html. Have fun!
Thank you for your wonderful simulation, can you please make a simulation for rotary drum dryer.
thats a wonderful tutorial. by the way is it possible to find the stresses acing on the hopper walls and calculate wall thickness. can you make a tutorial of it ?
Hey ajay, thanks for your questions. Indeed it is just a small modification to also dump the acting forces for the post processing. By this you can calculate the necessary wall thickness.
We will have it in mind and try to incorporate this explanation in upcomming tutorials.
Thanks for the great tutorial, I was wondering if it would be possible for you to explain how to install/ set up the LIGGGHTS code?
Hey Rhys, thanks for your question. In the next videos we will also make some tutorial how to install the software we use for our dem and cfd tutorials
Hello EngineerDo team,
I am a new user of liggghts and I am using lights to simulate granular particle interaction. I have figured out how to have a cohesive force and add a charge to specific atoms and types of atoms. However, I want to simulate the particles having electrostatic interactions with each other (a repulsive or attractive force). I have been able to have the particles interact with an electric field via the command efield, but I can not get the particles to have electrostatic interactions between each other. Any helpful hints, or do you know if liggghts can model this?
Thank you for your time
Thank you for your amazing tutorial. In paraview, is it possible to export a specific step as stl or another file so it can be loaded in Fluent? I tried export data, but there is a problem with the powder pack.
Hey Mohammad, you can export all STL files as new STL files. However, I fear you will not be able to export the particles as a STL surface. You can export the results to blender and from blender you can export STL. This would be possible. If you want to simulate the penetration by a fluid try to install the coupling using openFoam that we explained in a different video. This will help you to get rid of Ansys and do a direct coupling of particles and fluid.
Hey, super useful tutorial! I just noticed that the Young modulus is in MPa(N/mm2), it should't be in P(N/m)?
Yes you are right. The unit system is Si. So the pressure is Pa or N/mm².
Thank you for the tutorial!
Could you answer a question? At 15:00 of the video you explained the Generation and Insertion of particles. What does 10487 mean? How do you get this value?
Hey Kesley, thanks for your feedback.
This is a necessity of programming in LIGGGHTS. From pratical point of view it is important to know that this number must be a prime number and larger than 10000. So you could also replace the value for example by this numbers: 447883 447893 447901 447907 447943 447961 447983 447991 and the script works also.
The Background of the seed:
The value serves as a random number for the random generation of the particles within the generation region. By changing the value, the particles are randomly arranged differently in the generation.
Cheers
Thank you for tje tuterial can you make a video for coating tablets and modeling the spray droplets to calculate the tablets thickness after coating
That’s a very challenging task
Thanks for a very informative tutorial for beginners like me! I do have a question. Starting at 14:59 you are setting the particles. For a command like "fix pts1" and "fix pts2" there are the numbers 10487 and 11887, respectively. What do those numbers mean? There is also a large number in the "fix pdd1" command just below it.
Hey Jeff, these are prime numbers. Used by the randomisation algorithm. If changed you should generate a different particle distribution. Different in location, not in size. If this makes any sense. Like in reality you can generate with these a stochastic distribution from the same simulation experiment if repeated with different numbers.
@@engineerdo6568 Okay, so they are random number seeds. That is great to know. I appreciate the quick response.
Hi ! Thanks for the great tutorial to break down all the codes line by line. I have a question regarding some random numbers which are mentioned in the section of Generation and Insertion of the particles in this script. Also you have not discussed what they are in this video. Those are:
particletemplate/sphere 10487
particletemplate/sphere 11887
particledistribution/discrete 32452867
Are those numbers random or they have to be carefully specified?
Hey Taosif, thanks for your feedback.
This is a necessity of programming in LIGGGHTS. From pratical point of view it is important to know that this number must be a prime number and larger than 10000. So you could also replace the value for example by this numbers: 447883 447893 447901 447907 447943 447961 447983 447991 and the script works also.
The Background of the seed:
The value serves as a random number for the random generation of the particles within the generation region. By changing the value, the particles are randomly arranged differently in the generation.
Cheers
Hey, man! Thanks for your amazing videos! Do you have any tutorial on how to use liggghts on Windows?
Thanks mate. The usage is basically the same. I think you need a different tool for multi core simulations like mpiexec. The commands beside this, are the same
How do you manage to open .liggghts in Paraview
Nice tutorial.
Can you run liggghts on windows?
I'm confusing with one point: Don't you need a mesh between the silo and the ground?
Hey Andre, DEM is a meshless method. No need of a mesh like in CFD. You only need a mesh for every wall or object in the simulation. Running liggghts on windows is technica possible but we don't have a tutorial for that so far.
Thanks for the video.
I'd like to ask about choosing the timestep, what is the best timestep if simulating soil particles (especially clay particles)? Where can I read about this topic? I think I should try different timesteps until the results converge is this right?
Hey Sherif, thanks for your comment.
The selection of timestep can besometimes challenging. As a rule of thumb it is often recommended to use 10% to 20% of raygleigh time. But in the end it depens also on your particle system, the dynamics and the results you want to analyze. Hence, we can also often use 50% without changing the flow behaviour.
You can check your rayleight time percentage in liggghts by:
www.cfdem.com/media/DEM/docu/fix_check_timestep_gran.html
Cheers, Engineerdo
Thank you very much for the tutorial... Got to know a lot of new stuff.
I am using ParaView 5.8.0 and there's something wrong with the 'color' feature.
I was following this tutorial, I didn't get the whole list like you got for the 'particle_' , so I couldn't set the color to 'Velocity'.
Actually anything other than 'Solid Color' is unavailable. Can you please tell me what should be done to get it right?
Thanks in advance.
Hey Gourav, thanks for your question. You need to install/ load a plugin in paraview to get the additional information visualised. I think it is just updated for Paraview 5.4.1. Hence, I am not sure if it works for Paraview 5.8.0. too:
Here you can find further explanations and the plugin to download:
www.cfdem.com/paraview-plugin
github.com/richti83/ParaView_Reader_for_LIGGGHTS/tree/master/pre_compiled/Windows/5.4.1
Then you have to do:
1. Start Paraview
2. Select Tools -> Manage Plugins -> Load New
3. Select the liggghts reader file
Hopefully it will work. Please give us a feedback
Hello Engineerdo team
Is it possible to customise the number of particles I want to insert?
Suppose let's say i only want to have exactly 500 particles?
Excellent tutorial btw. Thank you very much
Thank you very much for the tutorial. This will be the first time I'll work with LIGGGHTS. It's been a great initial video. Also, would it be possible to add a solid object moving inside the grains, let's say, in a 2D geometry?
Hey Douglas, thanks for your feedback! It depends what kind of solid object you want to mimic. You could use superquaric for example to mimic a cube. In the public version the 6 degree of freedom feature is not implemented, which allows you to also enable free movement of cad bodies.
Cheers Engineerdo
@@engineerdo6568 thanks for the quick reply. Oh I see. I think I'll have to mimic a cylinder moving inside the grains. In that case, would I also have to use superquadrics?
Yeah I think if you just want to mimic a quite simple geometry like a cylinder it should be possible to insert the cylinder as superquadric and simulate the other particles as spheres. An Alternative would be to use coupled simulation, but this could be far too much depending on your desire.
Cheers, Engineerdo
@@engineerdo6568 it feels right to me. Well, again, thanks for your answers and tutorials, they were very informative. Have you thought about creating a complete course on LIGGGHTS simulations?
@@engineerdo6568 que no me a y M noauricio y papá y dando cuenta de la escuelay no me gusta cuenta de que no me gusta verla asi que es eso de la casa de mi mamá prestado a la casa en la casa de mi mamá prestado
Thanks for the nice video. I was wondering if you have video regarding LIGGGHTS installation on Windows-10
Thanks for your comment Shubham, no we don't have one. Maybe we do it on some point in the future. However, the simulation using linux makes more sense because it will be 10 - 20 % faster.
Hi. Thanks a lot for your great tutorials! could you put a uniaxial compression tutorial ???
Will be next in line
Hello There ! wonderful tutorial that you did there. can you make a video of combining 2 systems like hopper/silo and a screw auger(screw conveyor) ?
thanks in advance
Hey Nithin, thanks for your question. This sounds like a good example. Do you have any special geometry / applications in mind or can you even provide a geometry? Than it will be faster to do.
@@engineerdo6568 take aquaculture for example. You have fish food of typical dia of 5 mm (spherical) is held in hopper and it is carried to the outlet via a screw auger (constant pitch or variable pitch ) . You can take like 600 kg of fish food for pellets . Also interesting would be to conduct a stress analysis on hopper walls. Framing practical questions excites me 😅
@@nithinsivakumar2726 thats a very good example. Indeed this is possible to simulate. If the parameters for the simulation are set correctly - the bulk material must be calibrated - the simulation will show very accurate results.
This simulation can show not only stress on the wall but also the torque which is required to drive the screw.
Is this an application you are working on?
@@engineerdo6568 not yet , but I will in a couple of months as my thesis project.
Hi. Thanks a lot for your tutorials!! They are very useful. I would like to ask you a quick question. Is it possible to impose the motion of the silo lid by using a external file, containing X-Y-Z positions and orientations as function of time? Thanks in advance. Cheers!
Hi Jose, I have done something like for my MSc with regards to a vibrating screen. I wrote an Excel VBA which took the data and developed the LIGGGHTS script for the movement
@@mishalmohanlal2096 Hello. I just saw your response. Would it be possible to have a look at your thesis? Thanks in advance!!
@@josegegas I can’t share my thesis at this point, but happy to help out
ua-cam.com/play/PLSb6Nuu8dgCIr24ztBAOQPR0HUZoR2Fd1.html
Well done. One question, why not develope a GUI application to input those parameters to be set or edit automatically into the liggghts script? From where to download the windows version of LIGGGHTS?
Thanks
Hello, very good tutorial.
In my case I am currently starting to practice with LIGGGHTS-WITH-BONDS, and I use lpp to convert files to VTK format. At the moment of executing the input.liggghts file, I only get the 4 stl files in the post, but not the particles_liggghts file. I have tried using the lpp dump *. "Particle" command to get the file (this works for me with the built-in examples in liggghts), however I was unable to generate particles_liggghts. Any recommendation?
Thank you
Hey Bastian, you are welcome. If you like you can send the script to info@engineerdo.com and we will have quick look to your problem. Usually the way you have allready tried is the best way if problems like this happens: Take a liggghts example and modify it to your need.
Hey,
Thanks, for such a wonderful video. However, I am facing problem in visualizing the results.
I follow your tutorial every step. I am using Paraview 5.7.0 and when I show the results in ParaView, it is shown without particle movement and no particle appears. In addition, I am not able to change solid color of the particle.
Hey Yassin,
thanks for the feedback. First ensure you jump at least to 1 step.
For paraview, check also the comment of ence and us below:
ENCE 444
2 weeks ago
@EngineerDo Hey for the post processing tip, I believe you need the dump to vtk reader plugin. Viewers could change the particle dump portion of the script to:
dump dmpparticle all custom/vtk ${dumpstep} post/particles_*.vtk id type x y z vx vy vz fx fy fz radius mass
to output the vtk format.
EngineerDo
1 week ago (edited)
@ENCE 444 Hey Ence444, thanks for sharing your feedback. We appreciate.This would be also a good option to use the vtk format. We try to mention it.
Otherwise the plugin reader can be download and install via:
www.cfdem.com/paraview-plugin
Hi, thanks for the great tutorial. I followed everything until the last step and all worked out but the lid does not open when running the final animation.
Hey Vavek, in the lower part of the simulation script you will find a fix move/mesh command. This will move the lid. Please have a look to this command. I hope you find this helpful. Here is a link to the help: www.cfdem.com/media/DEM/docu/fix_move_mesh.html
Hi Engineerdo, it was a great tutorial.
By the way, how can i calculate residence time for particles?
Could you help me?
Hey Mohammad, thanks for your post. You could try to define a layer of particles for a type different of the other particles. Then you could track these particles during the postprocessing and calculate the time
First of all, thank you for doing an amazing job for showing how opensource software can be used for DEM simulations.
I encountered an issue: after downloading the tutorial files and trying to run it on Ubuntu, I get an error message that contact model is not in the whitelist. I cant put my head around that message. What does it mean?
Hey Mos-ta-fa, that means that this contact model is not officially supported. It can mean that it is calculating perhaps a bit slower. Just ignore it. It's not relevant.
Cheers, Hendrik
Hello friend, very grateful to your videos, with this tutorial I can already simulate my own models, but I have problems for paraview to show the particles of this tutorial, the solution I have found was to change the data output to vtk. Do you have any idea how to fix this problem in paraview and allow it to recognize the .liggghts format?
You must change the output format to vtk and the filename ending as well. You can check at the dump command in the handbook of liggghts which you can find in your folder structure
Hi, could you explain more on the scaling of the model? In the Silo.stl file the orifice of the silo is 400mm. If scaled down by a factor of 0.001, the opening will only be 0.4mm, which is smaller than the particle(radius of 20 and 30mm)? Thank you.
Hey Aldrich,
liggghts is allways interpreting the inuput stl as meters. Hence it would interpret the 400 mm as 400 meters. Hence, a scale of 0.001 is needed. Alternative you could export the stl from your cad program as meters.
Cheers, Engenieerdo
Hi EngineerDo, thank you very much. How can we convert .vtk file to .dat file or .txt file, I want to export the position, velocity(and other available parameters) of the particles with the time so that I can just plot it on excel. I have one more query, what are all the parameters do liggghts provide as output for the particles? Best wishes.
Hey Vimod, please have a look to the dump command www.cfdem.com/media/DEM/docu/dump.html Your command could look like this: "dump myDump all custom 1000 DumpFolder/vidmodsFiles_*.txt id type x y z vx vy vz radius mass"
Hey for the liggghts to vtk, use a custom/vtk, and change the particles_*.stl to particles_* .vtk
I'm getting an error 'Invalid dump style'
Hey EngineerDo, How did you opened .liggghts files in paraview?. They are not recognised in paraview in my PC.
Hey Venu,
The problem in paraview occurs due we used a special paraview reader. We will change it in general in the upcomming tutorials.
You can fix the problem by replacing the line in your script:
"dump dmpparticle all custom ${dumpstep} post/particles_*.liggghts id type x y z vx vy vz fx fy fz radius mass"
by this new line.
"dump dmpparticle all custom/vtk ${dumpstep} post/particles_*.vtk id type x y z vx vy vz fx fy fz radius mass"
Thank you for this Amazing tutorials, sorry can you suggest a free open source software for making geometry.
Hey Semmana, thanks for your question. There are some free options like student programs of commercial CAD Programs. Open Source solutions could be FreeCAD, Blender or Salome. None of them will work for problems in an industrial scale. However, for simulation in the context of a UNI they will fine.
Have fun!
Thank you again
Good afternoon, I hope you are very well, a query, how can I consider the damping in the hooke or hertz model in liggghts
Cheers
Hey Gerarda, the damping is implemented as the coefficient of restitution - the rebounce height. Look for "cor"
Hello, Can I extract the particles' location file after the filling process before starting discharge?
Can we also incorporate the deformation of boundary surface?
Hey Amar, thanks for your question. At all there are different ways to incorporate deformable surfaces. For example, it is possible to couple liggghts with a FEM program. Do you have any special application in mind?
Awesome tutorial! How can I do this simulation if I want to simulate continuous materials like clay instead of discrete particles falling out of the silo; is Liggghts still works for that or I would need something else? Thanks for your time.
Hey Jason, thanks for your feedback. You can also simulate cohesive materials like clay. For this you have to activate the cohesion of the particle in the model line. The simplest cohesion model in LIGGGHTS is the sjkr model. Check the manual for now:
www.cfdem.com/media/DEM/docu/gran_cohesion_sjkr2.html
In the tutorial it should work, if you change the follwing lines:
1.) After the rolling friction variable definition, you have to add the definition for your cohesion parameters, like:
####Cohesion Energy Density####
variable CED11 equal 100000 # particle particle cohesion
variable CED12 equal 12345# particle walll cohesion
variable CED21 equal 12345# particle wall cohesion
variable CED22 equal 0
2.) Add a material property like:
fix m6 all property/global cohesionEnergyDensity peratomtypepair ${natoms} ${CED11} ${CED12} ${CED21} ${CED22}
3.) Change the two model line command to:
pair_style gran model hertz tangential history rolling_friction epsd2 cohesion sjkr2 #contact model
fix walls all wall/gran model hertz tangential history rolling_friction epsd2 cohesion sjkr2 mesh n_meshes 4 meshes silo lid ground frame
I havent double check, but it should work. You can "play around" with the CED value to see how your particle will get more sticky, when increasing CED values.
@@engineerdo6568 Thank you very much, this really helped me a lot. I will try this out.
@@engineerdo6568 I am getting this error
(../contact_models.h:364)
ERROR: Unknown argument or wrong keyword order: 'cohesion' (../pair_gran_base.h:129)
*Update: I can get it to run after switching the order of cohesion sjkr2 and rolling_friction epsd2
pair_style gran model hertz tangential history cohesion sjkr2 rolling_friction epsd2 #contact model
fix walls all wall/gran model hertz tangential history cohesion sjkr2 rolling_friction epsd2 mesh n_meshes 4 meshes silo lid ground frame
However, it is taking much longer to run, is there a method to change the amount of particles or another way to shorten the run time?
@@jasonalt6232 The amount of particle is determined by the mass in your system and the grain size distribution. Using advanced models will increase the runtime. DEM Simulation for industrial problems can take weeks of computation.
Hence, you can decrease the total mass in your simulation or you can scale up the particle radii. Both will lead to smaller number of particles in you system.
@@jasonalt6232 Thank you this really helped me out! I was struggling with this for days!
Hi ! Thanks for your tutorials. I have a question regarding filling a packed bed. I have cylindrical geometry. I am trying to fill it with the default setup as given in the tutorial. After filling it with particles, I exported the position of particles as csv file. In csv file. the volume of the number of particles are way larger than the volume of packed bed itself. How is that even possible? Another question is that how can I filled the complete bed for a given porosity?
Hey thanks for the question Pratyush. I guess there is something wrong in the export command of the CSV. There should not be any change in volume
@@engineerdo6568 OK, I will check this out. However, please give answer of my second question. How can I filled the bed for a given porosity?
Thanks Again !!
Excellent tutorial. You made it so easy to start exploring this software.
I just have a small problem, when I want to change opacity to any other value that is not 1, the silo dissapears. How can I change that??
Hey David, that can be a bug of PARAVIEW or a problem with the graphics driver.
@@engineerdo6568 Thanks for the response, I decide to run Paraview in Windows where i have all drivers instead of Ubuntu and all went great. Thank you.
In input script, there is a pair coefficient when definition of contact model is described. What does it mean by?
There are some remaining of the underlying Programm lammps on which liggghts is based on. I assume that is some history code. Just don't worry about it
Hi guys! This tutorial is a very clear tutorial. Could it be possible to show how to run the simulations using the ubuntu terminal on windows (WSL)? It's the only way I've gotten liggghts installed on my computer haha
Hey Rafael, we are uploading a tutorial on the installation and execution of liggghts on ubuntu next week. I'm very sure that your questions will be answered.
@@engineerdo6568 Thanks so much!
Hey Rafael, the installation tutorial for Liggghts 3.8.0 is online. Please have a look: ua-cam.com/video/ru3119ozC6M/v-deo.html. Hope you find it useful.
Can you suggest me if I want to get a more detailed expalnation of the script, how can I proceed.
Hey Tushar, thanks for you comment. You can check the manual of liggghts: www.cfdem.com/media/DEM/docu/Manual.html
Here you will find all commands, we used in the script.
I am facing some problem when I am following your tutorial. Error: Fix inser/stream (id ins): inseration velocity projected on face normal is < 1e-3 .. Can you please provide any suggestion?
This is a very nice tutorial. But is there a way where we can learn LIGGGHTS step by step like we learn other software? Any online course or something like that?
Thanks Krishna, we are working on quite some new tutorial in different forms which will be available in the coming months. If you subscribe to our newsletter, you will get notified when we are ready...
Hey, thank you very much for the beautiful video you have made, it helps me a lot. Can you guide me to plotting or extract data from the thousands of dump.vtk files generated after the simulation by LIGGGHTS.
Hey DEEPAK, thanks for your comment. We appreciate. Have you checked how to load the files into paraview like shown in the end of the video? What exactly do you need?
Is there any way to enter a particle range in fix pts command?
Eg. Particle size ranging from 10micrometer to 100micrometer
Hey Asif, that's pretty easy. Just have a look to 14:58 in the video. You can use a grain size distribution by adding more than the two particle sizes in this video. There is no limit. Just copy the command and put in as many as you like
@@engineerdo6568 thank you.
Is there any way to find the volume of particle inserted in the container using the code or in paraview?
Its for finding porosity.
@@asifkamal7033 there is a way. You have to calculate the sum of the individual particle volumes. This leads you to the volume of the solid phase. The bulk material volume must be calculated in a hull volume. For a example you can use a clip filter to generate a box of particles out of you simulation. Then you have both volumes and you can calculate the void fraction.
Hello, can you help me with the reader of particle.liggghts i could not found it. Nice tutorial, thank for help me
Hey Jaoquin,
thanks for your feedback! The best way is to change one line in the liggghts script and re run the simulation:
The problem in paraview occurs due we used a special paraview reader. We will change it in general in the upcomming tutorials.
You can fix the problem by replacing the line in your script:
"dump dmpparticle all custom ${dumpstep} post/particles_*.liggghts id type x y z vx vy vz fx fy fz radius mass"
by this new line.
"dump dmpparticle all custom/vtk ${dumpstep} post/particles_*.vtk id type x y z vx vy vz fx fy fz radius mass"
Good luck, Engineerdo
Thank you for the nice tutorial. May I know how to run LIGGGHTS on Windows 10?
You can compile it on windows. I guess it will be pretty complicated
Great tutorial, I'm wondering if you can make a tutorial on how to installLIGGGTHS on windows.
Hey Cristian,
It is on our list.
Can I use the same procedure to simulate the binder jetting process?
thanks for the great tutorial,I am a rookie and I follow the video enter the code samely on my computer,but when I run the LIGGGHTS it reports error and shows that "mesh(id silo):all 283 mesh elements have been lost/left the domain.please use boundry m m m or scale/translate/rotate the mesh or change its dynamics" how can I fix this ? can anybody help me?
You must have changed something. One or more STL structures are not fitting in the simulation box. Please use paraview and look if the dimensions of the STL parts fit in your box
Do you have any idea, why i don't get the particles file in paraview? i did every step with your same files
Perhaps you created *.liggghts result files. Try to change the dump command from custom to custom/vtk and the filename to *.vtk. The result files are in the post folder.
Hey, is there a way to incorporate multisphere particles into this tutorial code?
Hey, you can incorporate multispheres in the example. Just have a look to
www.cfdem.com/media/DEM/docu/fix_particletemplate_multisphere.html
We will also consider multisphere in upcomming tutorials.
Cheers
Hello ,Engineer Do
I followed your whole tutorial but when i am trying to load files in paraview i am not able to load "particles.liggghts" as it is not supported by paraview.How do i fix this
Hey Saksham,
The problem in paraview occurs due we used a special paraview reader. We will change it in general in the upcomming tutorials.
You can fix the problem by replacing the line in your script:
"dump dmpparticle all custom ${dumpstep} post/particles_*.liggghts id type x y z vx vy vz fx fy fz radius mass"
by this new line.
"dump dmpparticle all custom/vtk ${dumpstep} post/particles_*.vtk id type x y z vx vy vz fx fy fz radius mass"
Cheers
@@engineerdo6568 Thank you for the solution.
Also is there any way I can the graph plot of your tutorial to check my results
@@saksham384 you can use the "plot selection over time" filter in paraview or write csv data from liggghts during the simulation.
Hello, how is the time step chosen in a suitable way?
As big as possible and as small as necssarey
hi how are you, do you have examples of the linear contact model using LIGGGHTS?
Hey Luis, to perform simulation using linear model, you can replace the command hertz by hooke in LIGGGHTS.
Additional you just have to add an additional command for the properties:
fix id all property/global characteristicVelocity scalar value
Have a look here:
www.cfdem.com/media/DEM/docu/gran_model_hooke.html?highlight=hooke
Cheers, Engineerdo
@@engineerdo6568 Thank you very much, it worked perfectly, for future doubts I will write to you soon, take care, thanks again
I created my own simulation for powder spreading for LBPF but it is giving me an error. "Substitution for illegal variable" What did I do wrong?
You are using a variable which was not defined before
Hi, I'm using inventor when creating the insertions face and I encountered this error: "Face defined as planar space is not planar", how to fix this ? Thanks
That's an annoying thing. Some meshes out of the CAD seem not to be orientated in that way, that every vortex is in the same plane.
You can use blender to bring them all in the same plane. Load the STL file in blender, then hit "s" to scale the geometry and the type in the axis perpendicular to the surface (e.g. X) the hit 0. The geometry will have no dimension in X direction, now, which makes it planar.
Good luck
@@engineerdo6568 Thank you for your reply. I will try it later. Also, do you have any suggestions to speed up the process of the simulation? I want to simulate around 420,000 particles, using the same physics as your example, and it took a very long time to be completed.
I had changed the dimension of solid model files but liggghts is showing all mesh files are missing
Hey Prince Goyal, thanks for your comment. What exactly is the error message you get? Have you changed the CAD model in a CAD model or scaled in the liggghts script?
If you are using larger solid model dimension, it is important, that you also scale the dimension of the simulations box (x y z), so the model will fit in.
Was this comment useful to you?
@@engineerdo6568 have decreased the dimensions and I changed the scale to 0.0001. it is showing error in lid and silo.it is showing that mesh have been lost or left the domain.
@@princegoyal1843 Hey Prince Goyal, I think I got you now, but sometimes debuggen without script might be challenging :)
If you dont change the dimensions of the simulation box (xyz) it works. Following Points may be helpful to check:
1.) Double check scale of alle elements:
STL:
fix silo all mesh/surface file Silo.stl type 2 scale 0.0001
fix lid all mesh/surface file Lid.stl type 2 scale 0.0001
fix ground all mesh/surface file Ground.stl type 2 scale 0.0001
fix frame all mesh/surface file Frame.stl type 2 scale 0.0001
Insertion:
fix ins_mesh all mesh/surface/planar file Insertionsface.stl type 1 scale 0.0001
2.) change the open velocity, so the lid will just open 0.04 m in 1 second, else the lid could move out of your box:
variable openvel equal 0.4/10
If you are also changing the dimensions of the simulation box to down scale, check:
3) the extrude lenght has been scaled down to 0.025, because otherwise the particle generation would be outside the box:
fix ins all insert/stream seed 86028157 distributiontemplate pdd1 &
mass ${fillmass} massrate ${fillmassrate} overlapcheck yes all_in yes vel constant 0 0 -0.75 &
insertion_face ins_mesh extrude_length 0.025
I hope it works now, else send the script to info@engineerdo.com and we will have a quick look.
Cheers
I would like to run it in my paraview but there is no plugin reader for that. How can I get it?
Hey Achmad,
The problem in paraview occurs due we used a special paraview reader. We will change it in general in the upcomming tutorials.
You can fix the problem by replacing the line in your script:
"dump dmpparticle all custom ${dumpstep} post/particles_*.liggghts id type x y z vx vy vz fx fy fz radius mass"
by this new line.
"dump dmpparticle all custom/vtk ${dumpstep} post/particles_*.vtk id type x y z vx vy vz fx fy fz radius mass"
Cheers
@@engineerdo6568 I am sorry, I asked this question before I read the whole comments. I fixed the problem by adding liggghts reader plugin into ParaView. I cant use vtk because my liggghts compiler doesn't have that feature. Cheers from Japan for a nice tutorial !!! Please keep this channel alive!
Best simulation, how I can install ligggth on windows 10? please
Hey Matteo,
A guide of install liggghts on windows 10 is on our list.
Cheers, Engineerdo
@@engineerdo6568 it's possibile insert on input file the properties of material function of temperature? For example volume function of temperature or heat specific?
Hey, Do LIGGGHTS have GPU support?
Hey venu, AFAIK LIGGGHTS is just compatible CPU
Cheers, Engineerdo
Appreciated, highly informative. I will follow the website for more tutorials.
Kindly DM me.
Thanks Dumisani
Hello EngineerDo team,
When i am trying to load the files in Paraview i get the message that reader for "particles_0.lights" could not be found, i am using Linux 20.4 - any ideas ?
Hey Mohammad,
The problem in paraview occurs due we used a special paraview reader. We will change it in general in the upcomming tutorials.
You can fix the problem by replacing the line in your script:
"dump dmpparticle all custom ${dumpstep} post/particles_*.liggghts id type x y z vx vy vz fx fy fz radius mass"
by this new line.
"dump dmpparticle all custom/vtk ${dumpstep} post/particles_*.vtk id type x y z vx vy vz fx fy fz radius mass"
EngineerDo thank you, I did exactly the same by comparing to the code from chute tutorial 👌🏼🙌🏼
Great, it works :)
EngineerDo thanks a lot I was missing “vtk”after “custom” 👍🏼
@@engineerdo6568 Hello team, did you work any script to run on parallelcluster using qsub ?
How can i run .liggghts on paraview?
HI. I would like to ask for help. When I am post-processing the data using Paraview, it does not include or read the group file particles_*.liggghts. Any idea what should I do? Thank you so much.
Please try to export you data as VTK. Then it will be possible to directly read into paraview
Thank you! It worked 💯
@@clintzonebatino9867 Very good!
plzzzzz give the link to install liggghts in windows plzzzzzzz
i am in stresss
Hey Saksham, a installation using windows is pretty complicated and you will see a decrease in performance. You can use our tutorial for ubuntu to install and compile everything you will need. The process will take less than half a day. ua-cam.com/video/ru3119ozC6M/v-deo.html
@@engineerdo6568 i did the same and made a same figure and all the same codding but when i run the code i am getting a error
"ERROR:Variable evalaution before simulation box is defined"
nice tutorial.
I am wondring how can we define the insersetion face (at 15.37 sec) when we do not load the geometry as .stl file. I understand that after the style mesh/surface/planar, i need to use fix (instead of file) followed by id of the fix that can create the insersion face surface. do i have to define it somewhere and if yes, how can i do it?
Hey Atul,
this command usually needs a mesh as stl to generate the particle. ALternative you can create the particles in a defined region, too. Therefore you have to define a region for particle creation and use the following command:
Have look to:
www.cfdem.com/media/DEM/docu/fix_insert_rate_region.html?highlight=region%20insert
Cheers, Engineerdo