openFoam Hangout on Air #2 - Cloud Simulation, presentation by SimScale
3 likes1,431 views
SimScale is a cloud-based simulation platform founded in 2011, aimed at providing accessible computer-aided engineering (CAE) through a web interface. It utilizes OpenFOAM as its backend for computational fluid dynamics and offers tools for pre-processing, visualization, and collaboration among users. The platform is designed for scalability, enabling users to perform simulations from anywhere with internet access.
openFoam Hangout on Air #2 - Cloud Simulation, presentation by SimScale
- 1. SimScale – Browser-based CAE. An online simulation platform in the cloud - including OpenFOAM. SimScale GmbH Lichtenbergstraße 8 85748 Garching bei München +49 (0) 160 97383156 1 www.simscale.de
- 2. Outline 1) Who are we? 2) The concept behind SimScale 3) Feature overview a) Pre-processing b) OpenFOAM solvers c) Visualization and Post-processing 4) Enabling users 5) Cloud and collaboration © SimScale GmbH, 2014 – www.simscale.de 2
- 3. 1. Who are we? The team behind SimScale. ● ● Founded in 2011 as a CAE consultancy Started by young TU Munich alumni: Mechanical engineering, Computer science, Mathematics ● Platform development started in 2012 ● Platform launch: August 2013 ● Based in Garching (near TUM campus) © SimScale GmbH, 2014 – www.simscale.de 3
- 4. The idea A browser-based platform for an easier access to CAE. CAD Upload Simulation © SimScale GmbH, 2014 – www.simscale.de Design decision 4
- 5. 2. The Concept A distributed software system with scalable infrastructure. Users Frontend ● ● Internet access & modern browser ● ● Completely web-based „Desktop look-andfeel“ JavaScript, WebGL, HTML CAE Backend Web Server ● ● Data management, storage & backup Load balancing and cloud management © SimScale GmbH, 2014 – www.simscale.de ● ● On-demand cloud computing resources Infrastructure as a service (IaaS) 5
- 6. 3.1 Pre-processing Geometry handling and meshing. ● Geometry workflow ● ● ● Format support (STEP, IGES, BREP, STL) In-browser 3D visualization Meshing ● 3D Tetrahedral workflow ● ● ● Fully automated Optional prismatic boundary layers snappyHexMesh ● ● ● Support for all types of refinements Automated for internal and external flow Quality control ● mesh clip ● checkMesh © SimScale GmbH, 2014 – www.simscale.de 6
- 7. 3.2 CFD Using OpenFOAM in the backend. ● OpenFOAM and SU2 are supported so far ● OpenFOAM is used as default for most applications ● Currently integrated functionality Incompressible icoFoam, pisoFoam, simpleFoam Turbulence laminar, k-epsilon, k-omega, k-omega-SST Multiphase interFoam, interDyMFoam Compressible rhoPimpleFoam, rhoSimpleFoam, rhoCentralFoam Conjugate heat transfer chtMultiRegionSimpleFoam Buoyancy buoyantPimpleFoam, buoyantSimpleFoam + Most other OpenFOAM solvers are accessible via case upload © SimScale GmbH, 2014 – www.simscale.de 7
- 8. 3.2 CFD Why OpenFOAM? OpenFOAM is a very mature framework for fluid mechanics ✔ Brings solvers for many technical problems ✔ Very large user community ✔ Widely used across various industries ✔ Well validated for many use cases ✔ Stringent and modular architecture ● Consistent input file format (dictionaries) → OpenFOAM is a sustainable choice for the CFD backend © SimScale GmbH, 2014 – www.simscale.de 8
- 9. 3.3 Visualization & Post-processing Different technologies for client side and remote rendering. ● Pre-processing: ● ● Client-side visualization for fast and intuitive interaction Remote Post-processing Rendering via WebGL Compute nodes Client ● Post-processing: ● ● ● SSL based on ParaViewWeb datasets are loaded remotely Displaying, interaction Data processing, Rendering Rendering via WebGL or image streaming (JPEG/PNG) © SimScale GmbH, 2014 – www.simscale.de 9
- 10. 4. Enabling users How can the access to OpenFOAM be simplified? ● The learning curve for OpenFOAM is somewhat steep ● OpenFOAM provides features which are very powerful but hard to use ● Multi regions (conjugate heat transfer) ● AMI ● snappyHexMesh (≳ 70 settings for a simple mesh) What can be done? → GUI & 3D interaction (e.g. for boundary condition creation) → Best-practice settings where applicable → Automated features → Library projects as template © SimScale GmbH, 2014 – www.simscale.de 10
- 11. 5. Collaboration New perspectives by using cloud computing. ● Central data management in the cloud → Data is accessible ● from anywhere ● at any time ● by different users ● Project sharing & public links ● Project library ● Advanced features for collaboration ● ● Interactive, real-time work on the same project Interactive post-processing © SimScale GmbH, 2014 – www.simscale.de 11
- 12. 6. Backend Infrastructure as a Service (IaaS) ensures scalability. ● ● ● ● SimScale owns almost no computing resources Computing jobs (meshing / solving) are prepared on the server and sent to the cloud provider. The user can specify the hardware on which the computation will be carried out Completely scalable, however slower due to multi tenancy and virtualization ● EC2 Specs ● ● ● ● Standard Instances up to 16 cores Arbitrary connection possible GPU acceleration possible 10 Gigabit Ethernet, No Infiniband Virtualization & Multi Tenancy effects → Speed down © SimScale GmbH, 2014 – www.simscale.de 12
- 13. Thank you. © SimAcc Engineering, 2012 – www.simucloud.de ©SimScale, 2014 – www.simscale.de 13