Latticept

Harlley Parno and I are writing a publication about shell and tube heat exchangers. All of this modeled and simulated in Dotmatics M-Star CFD using their GPU-native lattice Boltzmann solver. The image shows how close we can get to experimental results using a WM-ILES (wall modeled implicit large eddy simulation) based on the method of cumulants. This was calculated with just 16 million voxels. That's about 4GB of GPU RAM. You realize that means fully transient heat transfer simulation of shell and tube heat exchanger on a laptop? A LAPTOP! Pope says 80% of the turbulence should be explicitly resolved by the LES filter. Maybe that's required for external aerodynamic cases where pressure distribution over the aerodynamic surface determines if your plane flies or crashes. John Thomas and his team figured out a way to get very accurate results with just 30% of the total turbulent energy dissipation explicitly resolved by the subgrid filter. That means COARSE GRIDS! Low computational time! Getting stuff done! All on NVIDIA GPU's. #mstar #cfd #gpu #shellandtube #heatexchanger #coarsegrid

Simulate. ??? Produce. ?? Maintain. ??? What is the Boltzmann Reactor System? It's a system for maximizing your chemical production. Or how I describe it to non-technical executives: "Make the most money flow ?? out of your money machine ??." Latticept has completely re-envisioned the typical pipeline from pilot->production->reliability with a unified platform for chemical production. Start by simulating your exact chemical process using Dotmatics M-Star CFD that runs natively on NVIDIA GPU's. Incredible mechanistic physics from John Thomas team of computational physicists makes sure you can simulate damn near every industrial proess: heat transfer, agitation, reactions, multiphase flow, precipitation, mass transfer... even PID controls. Optimized process -> more chemicals. And the reactor? Well, the jacket has 30% more heat transfer than half-pipe jackets. More heat transfer -> more chemicals. And we've had a welding labor crisis for a long time, so I developed the reactor to be welding completely automatically. All of it. No welders -> faster lead times -> more chemicals. Now you invested in this incredible platform... How do we protect this? And how do we continue to make sure money flows out of the money machine? Well you inspect it. Every. single. weld. is butt welded and ultrasonically inspected. All of them. No unexpected outages -> more chemicals. Are you noticing a pattern? Contact me to learn more. #morechemicals #nvidia #hpc #gpu #mstar #cfd

I created a Custom GPT for Dotmatics M-Star CFD Documentation!?Use it here: https://lnkd.in/e2-Kz5q8 ? I've been crazy impressed with LLM's since OpenAI released ChatGPT.?You can make a 'Custom GPT' using commands to 'train' the responses. ? Use this to supplement your M-Star knowledge search as you learn to use the software. M-Star is FEATURE RICH and I routinely reference the documentation. ? So... why is it better than just.... reading the documentation like a normal nerd? ? Because you can just give it a screenshot and ask questions.?And it can complete tasks.?I gave it a screenshot from the "Create" menu.?These are building blocks of the GPU-native transient M-Star solver.?Each block you add executes in order. ? I then ask it to "explain each one and give an example of its use".?It goes one by one spitting out exactly what they do and how to use them.?Absolutely wild! ? Full output attached! ? #chatgpt #ai #mstar #gpu #hpc #openai

??TRANSIENT CO2 ADSORPTION?? OVER PACKED BEDS with Dotmatics M-Star CFD Here we simulate cleaning a flue gas stream using porous media, mass transfer, voxelized moving bodies and scalar transport. All solved on a single NVIDIA RTX4090. A DESKTOP! Latticept specializes in transient multiphysics digital twins of industrial processes using M-Star's novel lattice Boltzmann solver. Additional physics are easily overlaid on the computational grid with the fully transient solver. Here I use two of the new Voxelized Moving Body to form a rotary valve configuration. The bottom and top rotating spheres simulate opening and closing different flow paths to turn a batch process into an automatic one. As the bottom flue gas enters, it flows over the packed beds made of 13X molecular sieve. We use a porous media model and a UDF Region to adsorb (and desorb) CO2 based on local concentration gradients. (Programmed by our Harlley Parno, of course.) Carbon dioxide is desorbed via hot nitrogen and collected in a tail gas drum. You can watch the concentration darken as the concentration starts to increase. John Thomas and his team continue to add physics to tackle the world's hardest problems! Someone asked 'well whats novel about this new voxelized moving body? Probably nothing on its own... it's the combined capabilities of M-Star that are so incredible: DEM, heat transfer, interfacial mass transfer, multiphase flow, scalar transport and on and on. DM me to start a trial of M-Star. Don't want to buy the software? That's fine, I do consulting too! Already use M-Star and want this model? It's FREE! #gpu #hpc #gpucomputing #nvidia #mstar #cfd #co2

Check out this Pelton Turbine being rotated by the water jet! Dotmatics M-Star CFD has fluid-structure interaction with voxelized moving bodies and multiphase flow. Read that again -- bodies can move and rotate based on forces being applied by fluids. Using the novel lattice Boltzmann method that runs natively on GPU's, I solved this in less than one day on a desktop NVIDIA RTX4090. Think of the applications! M-Star Post is build on top of Kitware Inc. incredibly visualization toolkit to make beautiful renderings like this for your LinkedIn portfolio. I'm offering FREE 30 day trials including a virtual machine with a GPU on Amazon Web Services (AWS). That means you can remote into the trial and start using it immediately! DM me here or visit www.latticept.com to learn more. #mstar #cfd #kitware #nvidia #gpu #hpc

Latticept is excited to provide this massive volumetric dataset (1TB) for the #AI training community! I'm hoping the NVIDIA Modulus users of the world (like Michal Taká?, PhD.) can use this dataset for some fully transient 3D training of Fourier neural operators. Maybe Rishi Puri can glue a GNN to it?... Computed in Dotmatics M-Star CFD using the Lattice Boltzmann Method, this is a full volumetric dump of hydrofoil agitator starting from rest at different RPMs. All of it saved in Kitware Inc.'s Paraview format. The dataset is 1.2 million structured cells of a hydrofoil agitator in a pressure vessel starting from rest. It contains vorticity, strain rate vectors, velocity vectors, energy dissipation rate and pressure for 1 second of fluid flow. Timesteps are 5e-4 seconds. Plain LES turbulence is assumed for the subgrid turbulence model via eddy viscosity. RPMs included are from 5 to 60 are included in 5 RPM increments. Download here: https://lnkd.in/eiDzia_f Other finite volume solvers are showing off their AI-capabilities by "solving CFD in seconds". What they're really doing is spending weeks to generate data, then using it to train another model and expect you to be impressed when they click a button. Congrats? I guess? All of this was solved in just hours on a desktop RTX4090. Most of it is limited by hard drive saving -- it took me LONGER TO UPLOAD IT to AWS! DM me here or visit my website to to learn more about M-Star CFD! #ai #cfd #nvidia #modulus #kitware

Dotmatics M-Star CFD has a new feature: "Voxelized Moving Body" (VMB). To my knowledge, this implementation is entirely unique to M-Star and allows us to do some incredible stuff! Check out this simulation by the great Harlley Parno: absorption and desorption of CO2 through a filter via paired butterfly valve actuation. Now the VMB literally 'contains fluid' and stops the flow. Previously moving bodies could only be represented via the "Immersed Boundary Method". In this method the moving boundary is overlaid on top of the discretized volumetric space. As the boundary moves, the solver adds a momentum source to each voxel. This is incredibly fast compared to other moving mesh approaches, but there are some downfalls. The largest issue being the IMB cannot "hold fluid". Those fluid voxels are still there and the fluid "leaks" inside the moving boundary. Now when a VMB is occupying the space the voxels are literally removed. There's nowhere for the fluid to inhabit. M-Star is literally adding/removing voxels as the moving boundary moves through space -- allowing us to do unparalleled simulations in oil, gas and chemicals. John Thomas is always thinking of new features that can rapidly expand M-Star's footprint. Can you imagine doing this in ANy other SYStem with Moving or Deforming Mesh? On a desktop NVIDIA GPU? Contact me for a trial!

Dotmatics M-Star CFD is an explicit GPU-native multiphysics solver that runs on NVIDIA GPU's. All of the following are 'building blocks' of the simulation. Pay close attention to the 'Runtime' options. Because M-Star is an EXPLICIT simulation, we are marching through time with very tiny timesteps updating the variables as we go. P&ID loops? No problem. Not only can you model a dissolution process where particles literally convert into scalars in the liquid, but you can TUNE YOUR CONTROL SCHEME! On a desktop computer! Don't believe me? Bristol Myers Squibb literally did it: https://lnkd.in/ejfaWDdc #gpu #nvidia #fea #cfd #mstar

Recentemente, um cliente solicitou suporte para implementar um modelo cinético de adsor??o/dessor??o de CO? em filtros de carv?o ativado. Esse processo ocorre em duas etapas: primeiro, a corrente contendo CO? flui através do filtro, que adsorve o CO? até sua satura??o. Em seguida, as válvulas mudam de posi??o, permitindo a entrada de ar quente para remover o CO? adsorvido e assim iniciar um novo ciclo de adsor??o. O Dotmatics M-Star CFD implementou recentemente o novo corpo móvel voxelizado que, ao contrário do método da fronteira imersa (IBM), é impermeável, permitindo a simula??o de válvulas como essa. Representamos o filtro de carv?o ativado como um meio poroso e, utilizando outra nova implementa??o chamada de UDF Region, podemos realizar a transferência de massa somente dentro da regi?o do filtro (observe a concentra??o de CO2 adsorvido no filtro aumentando e diminuindo conforme alternamos as válvulas). Deseja melhorar os processos da sua indústria ou até mesmo projetar um equipamento especial? Entre em contato com a Latticept! #CFD #Engineering #Engenharia #Technology #Tecnologia #Industry #Industria #Innovation #Inovacao

SPRAY BREAKUP! Dotmatics M-Star CFD supports two-phase immiscible fluids and runs natively on NVIDIA desktop GPU's. This isn't just a VOF-type model -- the air and water are interacting to break the liquid blobs up coming out of the nozzle. M-Star also has Eulerian-Lagrangian hybridization... sometimes known as VOF->DEM. M-Star's implementation goes two-ways: the liquid becomes droplets and the droplets recombine to form liquid parcels. This makes sure we keep the mass balance in the system absolutely perfect. We can easily track the interfacial area between the two phases and can transport species between the two. Incredibly rapid multiphase mass transfer simulations on a desktop GPU... this solved in two hours!