Two decades after software vendors’ first efforts to put simulation and analysis capabilities in the hands of non-analyst engineers and designers, a wealth of new technological avenues have opened up to finally make this a practical reality.
By Bruce Jenkins, Ora Research
In the December 2016 issue of Design World, Part 1 of this series looked at how contemporary connections between CAD and CAE software environments have grown more sophisticated and intelligent than the industry’s early attempts. Now, we look at two more trends acting to democratize simulation.
— Meshless structural analysis and other recent technological advances that simplify and accelerate the traditionally time- and labor-intensive tasks of analysis model creation and problem setup.
— The cloud-based high-performance computing (HPC) revolution, driven by a wave of visionary young organizations and initiatives working to make cloud HPC-based simulation resources easily and affordably available to engineering organizations of all sizes.
De-bottlenecking model preparation and problem setup
A new generation of simulation software is bringing novel, powerful approaches to simplify the modeling and setup of analysis problems—a bottleneck where much of the time and expertise required to use sophisticated CAE have long been concentrated.
A leading example is SIMSOLID, a young software company whose structural analysis product of the same name was developed specifically for design engineers. Its most dramatic technological breakthrough is to let users perform analysis directly on imported CAD geometry without first having to perform the two most time-consuming and expertise-intensive steps in traditional FEA: geometry simplification and meshing. In addition, SIMSOLID can analyze complex parts and large assemblies not practical with traditional FEA, and do so efficiently on desktop-class computers, the company says.
Another such startup is simulationHub, a new, easy-to-use, cloud-based computational fluid dynamics (CFD) app for designers. Delivered as Software-as-a-Service (SaaS), the app is accessible from mobile, tablet and desktop devices. Intended to redefine the traditional CFD simulation process for complex CAD models, even working with dirty CAD assemblies that contain large openings, the app eliminates the CAD repair and de-featuring process from the CFD simulation work cycle. The company’s aim is to greatly simplify meshing, solve and post-processing operations to let designers focus on product design.
SimForDesign from Moscow-based FEA software developer Fidesys is a cloud-based CAE service that streamlines simulation for designers. Aimed specifically at designers, it does not assume its user to be a simulation expert. The software is meant to help designers understand structural performance of their designs, specifically displacements and stresses, to make smarter design decisions. Not meant to replace conventional CAE tools used by analysts for final design validation, SimForDesign is intended to make structural analysis an integral part of the design process by using automatic mesh generation and analysis to remove the complexities of setting up standard structural analysis of parts and small assemblies.
Pricing is on a prepaid-usage basis, starting with a $25 Standard package for one user, one compute hour, one GB storage, and five projects (plus free examples); users can then purchase additional credits as they use simulation more. Unlike with many services, SimForDesign users are charged only for compute time spent on geometry processing, meshing and analysis, not CPU core hours. The service automatically chooses the optimal hardware for each model to ensure best performance. The software can be deployed in local data centers if desired.
Cloud HPC revolution making simulation more accessible and affordable for all
The last two examples show how cloud computing is starting to reshape the simulation landscape, and for good reason. Simulation users have long been frustrated by constrained availability of HPC (high-performance computing) resources to run their computationally demanding applications. Expensive on-premise hardware could be hard to justify based on infrequent or fluctuating usage, while leasing time from supercomputing centers could incur exorbitant costs. But the explosive growth of commodity cloud computing in the past handful of years has completely rewritten this equation, making ultra-high-end processing power available and affordable for even the smallest engineering groups today.
Leading the charge to make cloud HPC resources readily accessible to engineering simulation software users is a new wave of young, visionary organizations and initiatives. One prominent example, San Francisco-based Rescale, offers software platforms and hardware infrastructure that let companies execute engineering and scientific simulations. Its goal, it says, is to “help transform stagnant, on-premise resources into an agile, optimized cloud HPC platform.”
Rescale’s cloud simulation and HPC platforms provide a range of software and hardware tools in one central location, giving engineers and scientists immediate and unlimited access to the exact resources they need on a pay-as-you-go basis. Rescale’s extensive list of software partners gives users turnkey access to more than 180 simulation software packages. Software pricing is either pay-as-you-go, or users can employ their own license server.
Another prominent example is SimScale, provider of a cloud-based CAE platform accessible entirely through a standard web browser that lets users seamlessly simulate, share and collaborate in its community of 65,000 engineering professionals. The company, headquartered in Munich, Germany, characterizes its mission as “harnessing the power of the cloud and cutting-edge simulation technology to build not just another simulation software but an ecosystem in which simulation functionality, content and people are brought together in one place, enabling them to build better products.”
Yet another example generating high interest and engagement is UberCloud, an online community and marketplace where engineers and scientists discover, try and buy computing power as a service, on demand. Engineers and scientists can explore and discuss how to use this computing power to solve their demanding problems, and to identify the roadblocks and solutions, with a crowd-sourcing approach, jointly with the UberCloud engineering and scientific community.
A fourth such initiative is AweSim. A partnership among the Ohio Supercomputer Center (OSC), simulation and engineering experts, and industry, its aim is to put simulation-driven design capabilities within reach of small to mid-sized manufacturers (SMMs). AweSim builds on OSC’s former Blue Collar Computing initiative to offer a new level of integration and commercialization of products and services for SMMs.