Cadence: A Natural Progression From EDA to CFD

We recently spoke with John Chawner, Senior Group Leader at Cadence about the company’s entry into computational fluid dynamics (CFD) technologies after many years of being known for its expertise in electronic design automation (EDA).

Cadence has long been a leader in electronic systems design, building upon more than 30 years of computational software expertise. The company applies its underlying Intelligent System Design strategy for creating innovative software, hardware, and IP solutions.

In turn, Cadence customers are some of  the world’s most innovative companies, delivering extraordinary electronic products from chips to boards to complete systems for the most dynamic market applications including hyperscale computing, 5G communications, automotive, mobile, aerospace, consumer, industrial, and healthcare.

Generally, computational fluid dynamics (CFD) is an aspect of multiphysics system analysis that simulates the behavior of fluids and their thermodynamic properties using numerical models. Cadence’s CFD suite includes application areas such as propulsion, aerodynamics, hydrodynamics, and combustion. What makes CFD platforms especially pivotal is their ability to adapt to specific instances of additional physical phenomena.

Cadence’s primary CFD products include:

Fidelity CFD Solution – Provides a streamlined CFD workflow for design, multidisciplinary analysis, and optimization in a single environment. The Cadence Fidelity CFD software is a scalable Cadence CFD solution that provides technology beyond the Omnis end-to-end CFD workflow and Pointwise legacy meshing. For over a decade, the CFD industry has recognized that progress in three vital areas has plateaued: numerical algorithms, modeling of turbulent and separated flow, and the exploitation of HPC assets. Fidelity CFD makes strides in all three; the two most unique aspects are its high-order solver and the first steps toward integrating the Omnis and Pointwise meshing technologies into a single, unique platform while also advancing them separately. The Fidelity high-order solver can provide 10 times the accuracy of classic flow solvers, and 3 times mesh speedup using the Pointwise solution within Fidelity CFD.

Fidelity Pointwise CFD Mesh Generation – Utilizes advanced mesh generation techniques as well as geometry model preparation capabilities for high fidelity characterization of fluids. It is well known that meshing is the most time-intensive part of the CFD workflow with one of the highest impacts, but is vital to ensuring high-accuracy and high-efficiency CFD simulations moving forward. Beyond meshing, however, Fidelity Pointwise can integrate into just about any CFD simulation workflow. Through its mesh export and import tools and pre-built relationships with other solvers and simulation tools, Fidelity Pointwise can enhance existing workflows. The entire team dedicated to Fidelity Pointwise is deeply connected to the greater meshing community and is involved in many aspects of the industry and influential guidelines like the CFD 2030 plan. 

A Calculated (and Smart) Leap of Faith Into CFD

After years of success and leadership in EDA, we asked why Cadence decided to formally enter the competitive CFD market. 

Chawner said, “There are actually three good reasons for that. Number one, we were kind of already there. Number two, it’s just different physics – the underpinnings are the same. And number three, it’s the customer pull. So, let me go through each of those in more detail.”

“About three years ago, Cadence started moving into multi-physics systems analysis, to move into the system space beyond the core EDA expertise in chips and boards – in other words, electronics.”

“Once you get into the multi-physics for performing electro-thermal co-simulation, you’ve already made that step into CFD. Keeping electronics cool provides forced air cooling, which is moving fluid, which is CFD. Once you’ve made that step and you’re utilizing computational fluid dynamics for that particular class of a problem, you realize there are a large number of other applications that you can apply it to.”

EDACafe Interviews John Chawner, Senior Group Director at Cadence about Computational Fluid Dynamics (CFD)

“That’s the first thing. We were already there, so it’s the next natural adjacency. The second thing was the physics –  Maxwell’s equations for electromagnetism versus the Navier-Stokes equations for fluids. Other than that, the underpinnings of computational software expertise are the same. Big complex problems with hundreds of millions, if not billions of elements that need to exploit high-performance computing (HPC) resources to solve those problems. The software infrastructure expertise that Cadence already has supports this new mission just for a new class of physics.”

“The third thing was the customer pull. A lot of the customers who we are currently serving either take the chips and boards that they design and put them in their own systems. And by system here I mean what Cadence and EDA people call a ‘big thing.’ If you call a chip and a board, a small thing, systems like planes, trains and automobiles are ‘big things.’”

“Those customers that are putting electronics into these big systems and many of them operate in a fluid environment. Planes, trains, automobiles incorporate systems for which CFD is critical to their design process. So whether they’re doing it themselves or they’re passing along those components to others who are doing the same thing, there is a market there to better serve our core customers simply by expanding our footprint into a different line of physics, which is computational fluid dynamics.”

New Research Into CFD

Chawner moved on to industry research and opportunities for computational fluid dynamics. 

“I come from the aerospace industry and for the last several years I’ve been involved in an ongoing cross-industry collaborative effort centered around a report that was written in 2014 called the NASA CFD Vision 2030 Study that laid out for NASA’s needs and the aerospace industry needs and what is required of CFD capability to get us to where we need to be in 2030.”

“Things like fully exploiting new next generation computational environments, HPC resources, GPUs, and quantum computing, as well as automation at all levels of the process. If you want to expand the usage and benefit of simulation to these other areas and maximize its impact, it has to be automated. The vision includes the fact that a single engineer has to be able to manage and interpret a large number of simulations in a time critical period like 24 hours without managing each one individually and extracting the engineering insights to a desired level of accuracy. All of those things can work together and drive computational science forward.”

“On the CFD side, in particular, we’re looking at spearheading advancements, especially the underlying numerics that support the computational techniques. There was a very interesting paper 15 years ago by a well-known Stanford CFD professor who made the assertion that the underlying numerics had plateaued a couple of decades previously and we’re still riding on algorithms developed in the ’80s and ’90s.”

Potential Fluid Flow Pathways in CFD. Determining what tools, equations, and models you want to use in CFD is an essential step in the overall simulation process.

“They work well enough, but they have known limitations that we could go beyond. That Stanford paper and the subsequent follow-up has spurred research efforts into probably the most challenging problem in all of fluid dynamics – predicting the turbulent flow of a fluid.”

“For example, turbulence is something we all experience on an airplane. The pilot announces to strap in your seat belt, and you feel that vibration. If you stick your hand outside of a car window and you hold it out flat everything is nice and smooth. Turn it, though, and you could feel that buffeting. It’s just the chaotic motion of the fluid. That is so difficult to compute and to compute accurately, that for many years it has been modeled in CFD software with supplemental equations that work well but are subject to certain bounds of applicability.”

“If we want to get outside that envelope and do better simulations, like when an aircraft is at a high angle of attack with its nose pointed way up, we need to be able to simulate things like turbulence – the transition from a smooth flow to that turbulent flow. There are techniques available to do that that we’re expanding into. Those kinds of things are in the background of just in general where CFD is going and what it means.”

How Fidelity CFD Software Supports Cadence’s CFD Goals

 “For Cadence, it’s a really exciting time. We announced Fidelity and launched it back in April of this year. It brings together all of the CFD assets in the entire company under one roof, under one framework, including the assets that were required, but NUMECA and Pointwise from last year, brings them all together and creates and has built a platform, an integrated platform that’s gonna be the basis for all of our CFD capability going forward into which we can plug and create the workflows that provide the automation that I mentioned earlier. The other thing that it does, and perhaps the crown jewel of the Fidelity line right now is we announced a next generation flow solver that features high order numerics and what we call scale resolving simulations as well as massive GPU acceleration.”

“We understand GPU acceleration because we know we have to compute as quickly as possible. Scale resolving goes back to that turbulence issue that I talked about before. Instead of the models, we are directly computing different scales of the turbulent eddies in those flows. For example, we have a technique that is also used by others, but it’s implemented in our flow solvers, it’s called Large Eddy Simulation, so we’re directly simulating some of those turbulent eddies to make it more accurate, and with that better accuracy, you get better predictions of heat transfer, you get better predictions of noise, aero-acoustics.”

“It means that we can make the fundamental elements of the calculations we’re doing smarter and more efficient, so we need fewer of them. So it’s an upgrade on the numerical algorithm that delivers accurate results, higher accurate results faster.”

Industry Verticals Benefitting from CFD 

“Virtually all industry verticals can benefit from CFD, but then there are some that, of course, you could imagine are at the leading edge of that. Aerospace and defense, for example, and aircraft – the ability to be able to shave fuel consumption is incredibly important to get drag correct. So at cruise conditions, you can use these techniques to do those kinda calculations, but also when a jet like that is landing at the airport and it has all the slats and flaps and landing gear deployed, that creates noise and to minimize the noise environment, we can use this high order solver with its scale resolving simulations to get the aero acoustics better so we can come up with mitigation strategies for noise.”

“Another place where we have a great deal of expertise are marine applications. We’re looking at surface ships, the marine industry has this great desire to, again, cut fuel consumption by shaping hull forms, by making the propulsors more efficient, but also those simulations extend to above the water line to do sea-keeping in rough seas to make sure the ship is stable, or to look at what they call wind-over-deck simulations, especially on the defense side of things, where you’re gonna be landing aircraft or helicopters on the aircraft.”

 “If you have a wind blowing over the superstructure, it creates all these vortices that can interfere with landing, and there’s been work done to actually populate flight simulators with data from simulation, rather than forcing the pilot to test it by actually going out to sea and trying to land on the aircraft carrier in different configurations. Another one is power generation, which is at the top of everyone’s mind right now, that’s what we would call turbomachinery – big things that spin. A boat is a big thing that floats, an airplane is a big thing that flies. But turbomachinery is pumps, compressors, fans, blowers, and then gas turbine engines like for the ones out on your aircraft as well, and that’ll be key to making those much more efficient in terms of their fuel performance and more robust in just terms of overall performance.”

“Automotive is an industry where you have not only noise considerations for passenger compartment cooling because one of the other big sources of noise is the fan that’s blowing to cool the electronics under the dash, so we have to minimize that noise. But also under the hood looking at the engine, whether it’s an internal combustion engine or an electric motor, its thermal environment has to be managed because regardless of whether you’re driving through Death Valley or going on a ski vacation in Vermont, that engine has to operate within its optimal thermal parameters. So being able to simulate that environment is key and will benefit greatly from the technology that we’re rolling out in Fidelity.”

Simulating turbulence and noise and how the air flows for vehicles is important, but it is also true for thermal simulations for electronics, if you want to maintain a certain temperature. 

Chawner said, “Yes, absolutely true. As the form factors get smaller for electronics, the ability to optimize that heat transfer using as small a fan as possible, or using passive cooling techniques, definitely, this will apply as well, and that’s really the crossover. Again, going back to the beginning, if you’re doing CFD for that, it makes sense to expand the footprint, and do it for these other applications as well, driven by what our customers want us to do and what they need us to do. So that’s what really makes it exciting.”

Final Thoughts on Cadence CFD 

It’s no secret that there are several players in the CFD, in the mechanical spaces, such as Dassault Systèmes and others. How does Cadence’s CFD solution stand up against these competitors that have been masters in their domain for many years? 

“Well, the good news is I don’t view this as a zero-sum game. With Cadence entering CFD we’re really driven by what our customers need us to do, both our existing customers from the EDA side who may need mechanical simulation support from the CFD side, and vice versa, our traditional mechanical CFD simulation customers who might need stuff from the EDA side. So the acquisitions of Pointwise and NUMECA brought to the table expertise in the marine applications, in turbomachinery, aerospace, and defense that we can leverage, again, specifically to serve those customer needs and then grow that footprint.”

“I don’t look at simulation as a competitive thing. I believe within the larger PLM arena, simulation is the fastest growing segment, behind only CAD. So this is a growing market where we have an opportunity to grow as well and serve our customers the best we can with this new technology as we roll it out and expand it and continue to develop it.”

 For More Information: Cadence

This entry was posted on Thursday, June 9th, 2022 at 1:39 pm. You can follow any responses to this entry through the RSS 2.0 feed. You can leave a response, or trackback from your own site.