Simulation Fidelity: Do You Need More, or Do You Want More? 

By Krystian Link 

I’m an automotive enthusiast.  And like all automotive enthusiasts, it’s hard to want just one fun car to own in the future.  For me, one of those cars might be a 2022 Camaro ZL1.  A friend of mine recently bought one and offered to give me a ride along, and he did not have to twist my arm.  This car has 650 horsepower and 650 pounds of torque.  For those reading this who aren’t car people, that’s somewhere between four to six times the power of your average car.  It was a sports car designed for the racetrack after all.  It did everything well, it delivered an emotionally-charged experience, was irrational for everyday commuting, and I wanted it.  But it begs the question, do I really need that much horsepower to achieve my objective of fun?  Can I choose a more entry-level sports car that might be cheaper, more balanced, or otherwise more practical like a Subaru BRZ?  Probably.  Likewise, simulation engineers want more fidelity and accuracy…but do we really need it in every stage of product design?  Probably not. 

The Two Sports Cars Of Simulation: Ansys Flagship Tools and Ansys Discovery 

In the simulation world, we have flagship tools (such as Ansys Fluent, Ansys Mechanical, HFSS, etc.) and we have design concepting tools like Ansys Discovery.  Ansys Discovery offers a unique value proposition: it’s cheaper and more accessible than an equivalent flagship tool, meaning a design engineer can learn it very quickly.  The sacrifice being Ansys Discovery’s simulation fidelity might not be as high when compared to the same situation using a flagship tool like Ansys Fluent or Mechanical.  Think of it as the entry-level sports car in my example above: it will get you 80% of the way there.  The flagship tools, much like the Camaro ZL1, is meant to get you the remaining 20%.   

When 80% Is Enough: Ansys Discovery in the Design Concepting Phase 

When discussing Ansys Discovery with some customers, the first question they ask is “what accuracy am I giving up?”.  A rational question, but it’s born out of fear rather than necessity.  That’s the point of simulation after all: more data everywhere.  But within the product development cycle, accuracy isn’t the only item of concern.  Companies want to launch the most optimal product as quickly as possible and as cost-efficiently as possible into the market.  That means not over-deploying some of its most costly engineering assets (like simulation engineers) on concept ideation activities typically handled by design engineers.  When a design engineer is trying to determine which concept design of five (or more) has the highest likelihood of making it to production, “close” is in fact good enough. 

“But what about accuracy!  We need to know what the design’s maximum tensile strength is to within 5 Pascals or what it’s temperature limit is within 3 degrees Celsius,” I hear you say.  Pause.  When you are at the concept phase of a design, do you really need certification/go-to-market levels of data accuracy?  Be honest.  You probably don’t.  And for the time it takes to set up Ansys Discovery simulations to “knock out” 3 of those 5 designs, and move the product development forward into design phase, you probably couldn’t run more than a single simulation using flagship tools.  Concepting is about accuracy by volume: “fail fast to succeed sooner” as David Kelley, founder of IDEO, would say.   

When The Last 20% Matters: Flagship Tools for Design Validation 

When you are launching a product and need to ensure it’s going to launch on time with quality, “close” is not good enough.  When you’ve isolated a design that performs reasonably well in Ansys Discovery, and you are preparing for the final leg of development, you want to be using flagship tools to ensure the product will pass certifications and final physical tests.  The risk is high at this stage of product development, so you absolutely should spend the time and use the best-of-the-best simulation tools.  The time and investment required to set up and run a high-fidelity simulation is more than it is in Ansys Discovery, and rightfully so.  When properly deploying simulation during the validation stage of product development, you can potentially avoid any (or all) of the following: 

That is not a holistic list, but the take-home point is that these pains can span days, weeks, or even months, causing irreparable damage to the product and company.  “But this product needs to get out the door!  We can’t afford to wait on simulation,” I hear you say.  Again, pause.  Would you rather have a product on the market first, only to be returned the next day for a warranty claim, or would you rather have a product that is delayed a few days, and be an overwhelming success with your customer base?  Trust me, you want the high-end sports car of simulation here. 

So next time, before deciding what simulation tool to use, ask yourself: do I need incredibly high fidelity or can I use something that’s simpler and more versatile?  You’d be surprised how far 80% can get you in the design process without having to pull the race car out of the garage.