I'm a chilean civil engineering professor. I specialize in modeling for earthquake engineering analysis. I'm also a developer of OpenSees.
My channel is about showcasing my research, giving tips and insight to the civil engineering community and talking about OpenSees.
www.joseabell.com
Пікірлер
Very Cool..!!! Thank You for sharing .. Cheers :)
Hello. When I run "gmsh.open("mymodel.msh")", I get the error message : Info : I'm busy! Ask me that later... How can I fix that ? Thank you !
Hi, I'd like to get Rayleigh wave velocity from numerical simulation like this. The nodal x or y directional velocity could be the surface wave velocity? or Is it just particle velocity?
Ive been wanting to build a cluster of dual epyc 7742s to learn on and try to play around with data bots on the web to predict trends. Ive got 42 gpus i could throw into the leoop as well if i build my own cases
there is a difference between teaching and showing skils .
Hi Jose - Brilliant video - I used gmsh2opensees to create a model in OpenSeesPy. One question: Once in Openseespy I have since modified the model, e.g. created nodes and deleted elements etc. So now when you call to visualise the model, it pulls up the old model created in gmsh, rather than the updated one. Would it be feasible to modify this, so that a model imported using gmsh2opensees which is then later modified by the user, could also be visualised in gmsh?
I would love to see a detailed configuration vdo. I’m want to build a cluster that can keep adding new node as my work load grows.
Are these 2.5Gbps cables enough? I'm looking for building my own HPC but i'm concerned about bottlenecks
Hola. Como podría generar una matriz sparse de esa malla? Y cómo podria hacer un grafo a partir de algunos puntos elementos de la malla? Gracias
scipy.sparse tiene todo lo que necesitas. Para grafos recomiendo networkx donde dependiendo lo que quieras puedes asignar cosas a los arcos y a los nodos del grafo.
Muchas gracias @@JoseAbell
Very Great !! I started in use of OpenSees with your tutorials. Great work.
Great to hear! Thanks for the feedback!
Thanks for the great tutorial! Any plan to make a video for compiling the parallel version of OpenSeesPy? The current cmake files seem to be set for the sequential python version only. 😆
Hi! That would be nice. I haven't tried out the parallel version of openseespy yet, but I'm aware that its there. I'll check it out.
Thank you very much!
You're welcome!
I have analyzed your Channel and found some problems. Do you want to talk about it for a while?
Thank you very much!
You're welcome!
How do i understand or observe the dispersion hehavior and higher mode with this simulation?
Not much dispersion going on here. Just numerical dispersion. You'd need more layers with sharper stiffness contrast to see more dispersion.
Thank you for your wonderful video. I have watch all your videos in your channel. I learned a lot. I had question whether you can do a video on QuadUp elements or 9 _4_QuadUP elements, or direct me on any sources that I can use them to build a soil model using both gmsh and openseepy.
Thanks for the feedback. Sorry it takes so long to respond. I have stuff going on. Anyway, at some point I want to do a review of many elements including UP solids. Can't promise anything soon though.
by the looks of I would assume you were a COMPUTER engineer, not a civil one 😂
I'm a huge all-around NERD!
Hi! the video is a couple months old! in the meantime there has been big changes on ChatGPT, especially with ChatGPT Plus. Do you still use it_
When i run example1.py i get this message OpenSeesError: See stderr output whats wrong with my openseespy
Is there more context in the error output?
Interesting to see the comparison! I feel that having the metadata is very helpful when sharing the data, but that is a great performance boost just using binary.
Performance in computing is always surprising. I am a firm believer in "seeing is believing" in anything regarding computer performance.
Amazing that this cluster probably has more performance than massive multi million dollar supercomputers from just a Decade ago
On a node-per node basis, yes, but as a whole certainly not. Supercomputers usually use slower clocks an rely on parallelism at scale to provide throughput.
Nice done. Can you put some video on a full 2D or 3D Frame like a building frame with codes?
Python esta buenicimo 😮😮
Excelente aporte muchísimas gracias
Hello, Dear sir, I hope this message finds you well. Thank you for making great videos and clear explanations. I am using Pycharm and I have installed gmsh but the comment box is not visible there. could you tell me which integrated development environment we should use and how, if it is possible make a video on how to install and how to use it.
Good idea for a video I think. I use Sublime Text 4 as my text editor for everything. I'll review my setup some time.
hi. Thanks. Is this available on the github?
Hi!. No, it isn't. What do you need specifically?
Es un lujo tener profesionales tan capacitados en esta plataforma dispuestos a compartir su enorme conocimiento de manera desinteresada. Muchas gracias José! Saludos desde Concepción!
Gracias por tus palabras!! Ando lento ultimamente con el tiempo como para hacer más. Pero ya vendrá.
I would be interested in an exemple of a modal analysis of a beam/plate
Neat, thanks
You bet!
Great simulation. Thnx
Glad you liked it
Thanks so much for sharing Jose. I am investing some time into learning Opensees with gmsh these days and your tutorials have been extremely helpful. I have been recently playing around with ShellMITC4 elements for modelling a pipe geometry and came to know that the “stresses” eleResponse outputs internal forces instead of stresses for that element type. The four node tetrahedral elements would output the element stresses when the “stresses” eleResponse command is called correct? On another note: I was wondering if you might be able to show how gmsh can be used to mesh a straight pipe geometry with four node tatrahedral elements and brick elements (e.g., 169 mm diameter, 4 mm wall thickness, 60 m long). I have been trying the ‘transfinite’ command which works well for the surface mesh but I am having trouble getting a good mesh for the 3D volume of the pipe.
Thanks for the feedback! I'm so glad this is helpful. For the ShellMITC4 "stresses" will output the internal "forces" integrated across the thickness of the layer (moments, axial forces, etc.) per unit length, as per plate theory. If you want stress at a given layer (if your're using layered sections) you need to request fiber output. I'll be doing a tutorial on shell analysis in the future. I'll be also glad to model a pipe with 3D elements, I might do a live this week and cover that in a "raw" format (no video edits, etc.) to save time.
Thanks a lot, I see you use union to merge two volumes, can you please advice on the following: if I have a large box (compututional domain) and I want to embed, say, a sphere inside it and keep the two volumes distinct but do a proper meshing, would you please provide few gmsh script commands to do so, thanks!
You can use BooleanFragments (or Coherence) to have a conforming mesh on the boundaries of geometrical entities, while still keeping the entities distinct. Check this out: SetFactory("OpenCASCADE"); Box(1) = {0, 0, 0, 1, 1, 1}; Sphere(2) = {0.5, 0.5, 0.5, 0.2, -Pi/2, Pi/2, 2*Pi}; BooleanFragments{ Volume{1}; Delete; }{ Volume{2}; Delete; }
Thanks for this video! Is using any nDMaterials, such as SAniSandMS or PDMY02 with TenNodeTetrahedron Element, possible or not?
Yes, it is! I think you'll see advantages in using high order elements with highly nonlinear materials such as those. Both in convergence rates and computational cost. TenNodeTetrahedron and the high order brick elements should afford similar advantages.
Thanks for clearly discussing and demonstrating with this simple example, the benefits of higher order mesh. Would look forward to your future videos..
Glad it was helpful! Thanks for the feedback!
Outstanding explanation, your clarity and depth are much appreciated. Keep up the excellent work!
And here, people, is the guy that helped me program the TenNodeTetrahedron element! Give this dude a follow! He's sure to do some great stuff.
Have u thought about using ubuntu server? Or u actually need a nice ui in the compute nodes? Also what are the gpus model?
Yeah I thought. I don't need UI on the compute nodes. Honestly I just wanted to get it done. I'll learn to do a network install for the next cluster.
Excellent! The reason for only using CPUs is due to the advantage in accessing a greater amount of memory compared to a GPU? Or is there another reason? Best of luck with the simulations!
It's mostly because opensees doesn't run on GPUs, too much branching on complex nonlinear finite-element simulations.
Genial. Disculpe Profesor, la Paralelización que se usa es nivel de ejecutar por ejemplo un modelo por cada nucleo, y asi poder correr varios modelos a la vez?. O por el contrario, a un solo modelo, descomponerlo y poder ejecutarlo de manera mas rapida?
Hola! Gracias por la pregunta! Ambas. En algunos casos queremos simular modelos "pequeños" miles de veces, por ejemplo para un estudio parametrico. Otras veces, simular modelos grandes que no caben en un solo computador.
nice, but does it run doom?
😅 i'll make a video running doom on it
Hi Jose, I hope this message finds you well. Your openseespy + gmsh tutorials are very helpful - thanks so much for creating and sharing these tutorials. I was wondering if you might be able to create a quick video covering how to use gmsh and openseespy to replicate the laterally loaded pile foundation example found in the opensees wiki examples page.
You're very welcome! Regarding that example, it is a 1-D pile with lateral springs. It kind of defeats the purpose of using gmsh which is more apt for continuum models. If you want to use openseespy, I'd suggest simply translating the TCL files to the python equivalent. ChatGPT would likely get most of the translation right. You could still visualize on gmsh your results after that. Regards!
@@JoseAbell Many thanks Jose. That is a good point. I am indeed mostly interested in seeing if gmsh could be used to visualize stresses or strains in 1-D beamColumn elements at fiber section integration points. I wonder, for example, if gmsh has the ability to show a visual rainbow color distribution of the stresses or strains along beamColumn elements. I will do some digging. I am just starting to play around with gmsh and gmsh2opensees, and your videos are very helpful in this endeavor - thanks again!
Thank you so much, I started to learn both of gmsh & python recently, the video is so helpful. I aim to use gmsh with python to model electromagnetic problems.
Glad it helped!
Hello sir if possible pls share vedios or link which include modeling of slab by using opensees..
Great idea! I'll prepare a video on this in the future. There's lots of details to address.
@@JoseAbell sir pls do as much as early.
How to make slab model by using openseespy pls make vedio which include step by step explanation of code..
Helli si rply pls..
I'm thinking of making my next video (next week) about this topic.
@@JoseAbell ok sir thank u. slab explanation should contain 2D and 3D .
This guy is posting homophobic content on twitter, just so you know
As an observant catholic I do oppose gender ideologies and am pretty vocal about it on twitter, come what may. I hate no one.
Bueno se intenta entender
Aquí intentamos explicar tambien. 😅
Interesante trabajo, imagino que la defensa de su proyecto fue exitosa. Saludos desde Mx.
Gracias Carlos. Esto es parte inicial de un proyecto de investigación. Son solo resultados preliminares para mostrarle a la gente en qué estoy trabajando en el momento.
@@JoseAbell ¿Cómo va el proyecto de la simulación de sismos ?
Así sería el terremoto
Al menos en cuanto lo que predicen las simulaciones.
Amazing! It would be great if could make another video on how to formulate element and global stiffness matrix.
Great suggestion! What elements are you thinking about?
Soon we will write our own FEA and CFD codes. Later, the IA will learn how to do FEA and CFD jobs using all the information, videos and equations on the Internet. Few years more and the IA will do theses analysis instantaneously without using a single equation or code.