English
Log in
English

FAQs

Frequently asked questions about Onscale’s software, plans, and pricing.

About

  • OnScale is the first Cloud Engineering Simulation platform. OnScale combines powerful multiphysics solver technology with the limitless compute power of cloud supercomputers. With OnScale, engineers can run massive numbers of full 3D multiphysics simulations in parallel to

    create true Digital Prototypes – digital representations of physical high-tech devices that capture the complete behavior of a device over its operating envelope. By shifting expensive and time- consuming physical prototyping to digital prototyping, OnScale massively reduces cost, risk, and time-to-market for R&D firms pushing the boundaries of new technology.

  • OnScale is the combination of high-performance natively-parallel multiphysics solvers developed for the most demanding organizations in the world, public or private cloud supercomputers, a functional GUI, an API to efficiently integrate OnScale into any design work-flow, scripting languages to fully customize simulations, and plugins to enhance engineering capabilities. Data importing and exporting as well as postprocessing are included. Engineers can run simulations on high-performance computers in the cloud by letting the software do all the heavy lifting. OnScale is multi-tenant - there are no software licenses and no IT license management issues. OnScale comes with world-class data security and uses best-in-class security components from our cloud partners, AWS and GCP. Simply create an account and start simulating.

  • Computer-Aided Engineering (CAE) is the use of computer software to perform engineering analysis. The computer software is used to model, simulate, analyze and optimize designs, products, and processes. For engineering , there are usually three steps: preprocess, solve, and postprocess.

     

    Preprocessing is the generation of the design geometry and the definition of operational factors to be applied to the model such as loads and boundary conditions. Preprocessing also includes the geometrical discretization, or meshing, of the design or product under investigation.

     

    Solving is the appropriate numerical discretization of the mathematical model representing the design or product being analyzed. Solving also includes the numerical solution of the resulting numerical discretization.

     

    Postprocessing is the visualization and derivation of figures of merit from simulation results. Common values of interest are maximum, minimum, average, isocontour, integral and derivative of simulation results along keypoints, edges, surfaces and volumes. Visualization includes the representation of simulation data on keypoints, slices, surfaces. Vector fields and streamlines are also included.

  • The purpose of Computer-Aided Design (CAD) is to create a digital geometric representation of a physical design or product. With Computer-Aided Engineering (CAE) the physics characteristics of a design or product, like material composition and operational conditions, are represented mathematically through a set of partial differential equations (PDEs), which are then solved with numerical techniques. While CAD allows designers and engineers to define the geometrical features of a design or product, CAE allows the digital analysis of its behavior when subjected to expected initial conditions and loads.

  • CAE applications span a wide range of analyses, designs, products, and industries. A few examples include:

     

    ● Stress and dynamic analysis on components and assemblies

    ● Thermal analysis for electronics packaging

    ● Fluid flow analysis through Computational Fluid Dynamics (CFD)

    ● Acoustics analysis

    ● Mechanical event simulation (MES)

    ● Drop testing simulation

    ● Micro-electrical-mechanical systems (MEMS) design

    ● Non-destructive testing (NDT)

    ● RF Filters design

    ● Sonic and ultrasonic flow meters design

    ● Acoustic and ultrasonic transducers analysis

    ● Optimization of products or processes

  • Multiphysics simulation is the mathematical modeling and numerical solution of the behavior and interaction of physical phenomena simultaneously interacting, or coupled, with each other. Real-world designs, products, and processes are inherently multiphysics. For example, a car exhibits a variety of coupled physics across multiple components. If we consider the engine, the injection of fuel in the cylinder, its compression by the pistons, its ignition forcing the piston down for another stroke, and the release of exhaust fumes includes the following coupled phenomena: fluid flow, heat transfer, structural mechanics, chemistry, and acoustics. While approximating the interaction between all physics involved may result in useful preliminary insights, only multiphysics simulation can provide an accurate representation of such a system.

  • Transient thermal analysis calculates temperatures and heat fluxes in a product or system over a specific time range and for given material properties, initial conditions, and boundary conditions. Engineers commonly use the temperature field that a transient thermal analysis provides as input to structural analyses for thermal stress evaluations. This is also an example of multiphysics analysis. The coupling between the heat transfer and structural mechanics physics can one-way (weak) or two-way (strong) depending on material properties and geometrical characteristics.

FEA/FEM

  • The Finite Element Method (FEM) is the numerical technique used to solve the Partial Differential Equations (PDEs) to be solved in a Finite Element Analysis (FEA). With this method, the geometry or spatial domain is divided up or meshed into smaller finite elements where the mathematical solution is approximated with functions that can be, for example, linear or quadratic. When all approximated solutions are assembled together to represent the whole domain, the resulting system of algebraic equations is solved with dedicated numerical algorithms. The most commonly used are multigrid solvers and frontal solvers.

  • Finite Element Analysis (FEA) is the engineering analysis of designs or products whose behavior is simulated using the Finite Element Method (FEM). The physics characteristics and behavior of a design or product are represented mathematically through a set of Partial Differential Equations (PDEs), which are then solved numerically with FEM.

  • Computer-Aided Engineering (CAE) is the broad term for the use of computer software for the digital analysis of engineering applications. Finite  Element Analysis (FEA) is a type of CAE-based engineering analysis using the Finite Element Method (FEM).

  • A Finite Element Analysis (FEA) solver includes a preprocessor, a numerical solver, and a postprocessor. The preprocessor produces the geometrical and mathematical data representing the physical design or product being analyzed. The solver performs the necessary numerical calculations to solve the mathematical problem under investigation. The solver can be for example static, transient, eigenvalue, linear, or nonlinear. The postprocessor visualizes the numerical results obtained from the solver and allows for the derivation of value of interest such as maximum, minimum, and flux through a surface.

  • In Finite Element Analysis (FEA) convergence means that the solution of the Finite Element Method (FEM) approximation approaches the true solution of the Partial Differential Equations (PDEs) as the geometry, or spatial domain, is divided up or meshed into smaller elements. In summary, convergence refers to the impact of the mesh refinement on the accuracy of the solution.

     

    Typically, the smaller the mesh size, the more accurate the solution as the behavior of the design or product is better sampled across its physical domain. The higher the accuracy, the larger the simulations can become in terms of data to store and handle, which translates to a longer run time. Engineers often perform convergence studies, i.e. mesh refinement vs. accuracy, to obtain the optimal balance between accuracy and run time.

  • In a Finite Element Analysis (FEA) performed in the time domain, the time step is the time increments for which the Partial Differential Equations (PDEs) governing the problem under investigation are being solved. When explicit numerical solvers are used, the largest possible value of the time step is a function of the time it takes for the solution to propagate within the finite elements. Such value depends, among other factors, also on material properties.

  •  There are many different types of finite elements used in FEA such as:

    ●  1D, 2D, and 3D.

    ● Linear, quadratics, and cubic.

    ● For structural analysis the most common are: truss, beam, shell, membrane, solid.

    ● Truss elements for modeling line-like structures that support loading along their axis.

     

    - Beam elements for slender structures which resists twisting and bending.

    - Shell elements for structures which have a significantly smaller dimension, e.g. thickness, than the other dimensions.

    - Membrane elements for thin fabric-like surfaces.

     

    ● Connector elements for applying a desired behaviour between two elements, e.g. connect them through a spring.

    ● Infinite elements for unbounded domains.

  • Preprocessing in Finite Element Analysis (FEA) includes: generation of the design or product geometry, discretization or meshing of the geometry, application of physics, material properties, boundary conditions, and initial conditions. Preprocessing produces the mathematical data needed by the following solving step.

Getting Started

  • Head to www.onscale.com and click on login to Create New User account. New users can enter their details in all required fields to complete the sign up process. Please contact sales@onscale.com to active you 10 Core-Hour per Month Free Account

  • The Core Hour balance is shown in the accounts page.

     

    ● Head to www.onscale.com
    ● Log in with your credentials
    ● The account’s Core Hour balance will be visible

  • You can get support through the following methods.

     

    ● Search our Help Center for answers

    ● Post in our forum

    ● Submit a support request

Payments & Billing

  • Free Subscription — is ideal for a single engineer or collaborators working on small projects. Free Subscription gives users 10 Core-Hours/month and the ability to purchase additional core hours for $4/Core-Hour On-Demand.

     

    Team Subscription — is suitable for small engineering teams. This package provides 300 Core-Hours/month or 4,000 Core-Hours/year and the ability to purchase additional core hours On-Demand. For more information and to request a quote, please contact sales@onscale.com.

     

    OnScale for Enterprise Customers

    Our Enterprise package is a custom package for large engineering teams. It starts at 1,600 Core-Hours/month or 20,000 Core-Hours/year and goes up from there. The bigger the Core-Hour commitment, the lower the cost per Core-Hour. For more information and to request quote, please contact sales@onscale.com.

     

    OnScale for Academic Customers

    OnScale is proud to support engineering research and education! We offer Free Core-Hours for university students and greatly discounted Academic Core-Hours to university researches and instructors.

    We recommend starting with our Free Subscription with 10 Core-Hours/month and no limitations on the software, users or features.

    For additional simulation capacity, we offer Academic pricing for Team Subscriptions. For more information on Academic pricing, please contact sales@onscale.com.

  • For Team Subscriptions - OnScale accepts the following payment methods: Credit Cards (Visa, MasterCard, American Express, Discover, JCB, Diners Club), Purchase orders.

     

    For Enterprise Subscriptions - OnScale accepts Credit Cards and Purchase orders. All Enterprise Subscriptions are paid annually in advance.

  • If you received an error while we attempted to charge your credit card, please double-check your payment information and try again.

     

    You will need to contact your bank for the reason why the card was declined. Here are a few of the most common reasons why cards get declined:

     

    ● Incorrect credit card details (name on the card, credit card number, expiration date, CVV)

    ● Insufficient funds;

    ● Some banks will reject charges based on location;

    ● Some banks will reject charges based on their own fraud rules;

What is The Cloud

  • The cloud is a server computer located in Amazon AWS server farm at a remote location in which OnScale solver is installed.

  • (Where does my job go?) Your simulation input file is encrypted to ensure a high data protection and sent to Amazon AWS sever where the solver process and calculate your results. Those results are then saved in the cloud and need to be downloaded back to your local computer.

  • As your simulation is done on the Amazon AWS cloud server, the results are saved there by OnScale solve in a special encrypted location. You can access to this location with your OnScale account and download them on your computer by clicking through the "storage" window in OnScale.

  • When you launch an analysis with OnScale Cloud Solver, you spend from your reserve of Core Hours. The estimator is a small program which calculate how much time the first step of your analysis takes and provide with an estimation of the final analysis time and the Core Hours required to run the simulation.

  • When you do simulation, you are hundreds of parameters that you can set and which have all an influence on the simulation results. The traditional way is to change a bit each of them and to see how the solution vary. If you do that one simulation after one another, you will have to run hundreds or thousands of simulations one after the other, monitor the parameter changes and the results, make some screen captures and curves, this can quickly transform itself into a huge task.

     

    OnScale allows to set up some parameters that can vary in between certain numerical bounds and "sweep" through those parameters to quickly launch on the cloud hundreds of simulations in parallel with just one parameter changed. Results can then be handled by batch using script and what is called the "design space of the solution" can be reconstructed!

    This method is much faster and efficient than the old sequential trial and error method and allows you to make real breakthrough in the design of high end systems designed, because you can quickly find through an optimization study the right combination of parameters.

  • In Cloud computing, a "job" is not what you would usually think it is. A job is a computing task that the server must execute and process.

     

    Typically, all the information required for a job are written within a text file (number of cores, memory, program to be use and links to input files) and this text file is sent to the remote cloud server where a "job" is created.

Ready to Get Started?