Changing CFD

Share this on social media:

Issue: 

Gemma Church investigates how CFD providers are lowering the barrier for simulation

Computational Fluid Dynamics (CFD) is a staple fluid mechanics technology, where numerical analysis and data structures are used to analyse and solve a host of fluid flow problems. This established solution is now branching out, tackling everything from environmental issues to heart conditions and aeroacoustics.

This diversity of applications has affected the design and functionality of traditional CFD software, where engineers can now take on diverse problem sets with simulation and modelling. Andy Fine, Altair’s vice president for CFD solutions, EMEA, said: ‘We are dealing with increasingly complex physics when looking at CFD problems. While most CFD technology is extremely mature, it’s very expensive and requires a lot of specialist knowledge. We are trying to lower these barriers to entry by providing increasingly accessible, cheaper solutions.’

The accuracy versus performance trade-off is another major challenge to democratise CFD, where some applications require large computational resources to meet user expectations. Ed Fontes, CTO at Comsol, explained: ‘Here, you often find a conflict between the user’s expectation and the software’s ability to deliver accurate results to a small computational cost. Unfortunately, there are no free lunches: Accuracy costs.’ 

Fontes added: ‘It is easy to get colourful plots at low computational costs by adding lots of artificial diffusion, but this yields inaccurate results. We invest a substantial portion of our development time in numerical methods and solvers, in order to solve problems accurately and on desktop computers.’

Multiphysics coupling is another increasingly important consideration in CFD, where tools must address usability issues. Fontes explained: ‘It should be straightforward to couple CFD to other types of analysis, for example structural mechanics (fluid-structure interaction), chemical reactions, heat transfer, electromagnetic fields, equation-based modeling, etc. The solver settings for CFD should work together with the solution for other numerical equations added by other phenomena, either fully coupled or using a segregated solution strategy.’

Comsol’s Application Builder is addressing these usability concerns. Fontes said: ‘It allows CFD experts to build dedicated graphical user interfaces on top of specific CFD problems defined in Comsol. These experts can build applications that can be used by a larger community of engineers and scientists that are domain experts, but may not be CFD experts.’ 

Comsol has also optimised many of its existing tools for CFD over the last few years, including its meshing tool, iterative solvers and tools for results evaluation. ‘The mesh and solver projects have been very important for CFD and are still large projects, so we are in no way done. We are proud of the ability to visualise and evaluate results,’ Fontes explained. 

Altair has recently increased its focus on CFD, targeting specific and specialist applications across multiple industries using a suite of software solutions. Fine explained: ‘One size does not fit all in the CFD space. We make a decision for each client, based on their application, to find the best solution.’

Altair runs a unique licensing system called Altair Units. The company’s full suite of software solutions is available via Altair Units from its marketplace Altair One, where engineers can also get access to specialist support teams, including a CFD division, and third-party solutions from the Altair Partner Alliance. Altair’s CFD simulation suite includes Altair AcuSolve, which allows engineers to tackle several general purpose CFD problems, including flow, heat transfer, turbulence, and non-Newtonian material analysis, for example. 

Altair ultraFluidX and Altair nanoFluidX are application-specific CFD solutions. Altair ultraFluidX was developed specifically for ultra-fast prediction of aerodynamic properties for a variety of vehicles, whereas Altair nanoFluidX is a particle-based hydrodynamics (SPH) tool to predict fluid flow around complex geometries under complicated motion. ‘These tools allow us to tackle problems that are practically impossible to work with using traditional CFD,’ Fine said. 

One of these emerging areas is aeroacoustics, where the noises and vibrations created by an object in motion are investigated and reduced. These sounds are now more noticeable than ever, thanks to the development of electric vehicles, which provide a much quieter ride, compared to traditional vehicles.

The field of aeroacoustics is also increasingly relevant in other industries, including machining and tooling where manufacturers want to reduce noise created by large cooling fans, for example. Fine said: ‘This is a key and unique application area where we have the domain knowledge and understanding to solve these issues, not just simulate them, but also predict corrections and geo-spatial modifications that can solve any number of problems.’

The fields of plasma science and electrochemistry have both expanded significantly in recent years, according to Fontes, who explained: ‘Here, we find problems involving CFD, electromagnetic fields, chemical, and electrochemical reactions, but also the usual conjugate heat transfer problems (coupled heat transfer in solids and fluids with CFD in the fluids) that emerge in all processes where heat is required or generated.’ 

Biomedical applications such as heart pumps and other medical implants are also ‘an interesting area of application,’ according to Fontes, who added: ‘Medtech is a fascinating new area for CFD. If we look at the traditional fields where simulation and modelling is often used, they are simple systems, compared to biological systems. Now, thanks to advances in CFD technology, we can deal with increasingly complex biological systems.’

Comsol Multiphysics was recently used to simulate the HeartMate 3 left ventricular assist device (LVAD) from Abbott Laboratories. In patients with a poorly functioning left ventricle, an LVAD pump is responsible for circulating oxygen-rich blood through the body. CFD and electromagnetics modelling helped the Abbott team design a powerful, efficient and hemocompatible pump with pulsatile flow, which more closely mimics a functioning heart.

Freddy Hansen, a physicist at Abbott Laboratories, said: ‘It is easy to enter your own mathematical expressions in Comsol, in particular as you can type in an equation anywhere you can enter a number, and for us this greatly simplifies comparing different designs for their hemocompatibility. To do this, we enter custom algebraic and differential equations and solve these alongside the CFD. We find that the relatively large gaps that we have in the HeartMate 3 between its rotor and the surrounding walls – gaps that are only possible with a magnetically levitated rotor – offer the design that is the most gentle on the blood.’

Such work could revolutionise the field of medicine, as Fontes concluded: ‘Medicine today is very much trial and error but, using CFD, we could analyse a model and then create treatments that are not only cheaper, but adapted to an individual’s symptoms and physical characteristics.’

Environmental technology is another growing CFD field ‘involving processes such as carbon sequestration and processes related to the hydrogen society,’ according to Fontes. ‘CFD and fluid-structure interaction of wind turbines has been a popular application for at least 10 years, but also the design around the wind turbines becomes interesting. For example, acoustics, water flow for offshore farms, and airflow around wind farms presents many interesting CFD and multiphysics problems involving, for example, fluid-structure interaction.’

Smokin’ CFD

The Barcelona Supercomputing Center (BSC) is currently working on the Estimate project, a European initiative from the Clean Sky Joint Undertaking. Researchers are developing advanced CFD software to help the design of the new generation of efficient and sustainable Ultra High Bypass Ratio engines, with a focus on soot emissions.

When at full power conditions, soot emissions tend to be low but, at intermediate conditions, particulate formation can be significant. Estimate uses a multidisciplinary approach to predict soot formation from chemistry oxidation to particle formation, based on high-fidelity simulations validated with reference experiments. 

The project will shed light onto the influence of the combustion model and the treatment of turbulence chemistry interactions on the prediction capabilities of the soot models, in terms of soot volume fraction, soot volume density and particle size distributions.

The team is currently developing these models, and is addressing a range of challenges in its work. Oriol Lehmkuhl, large-scale computational fluid dynamics group leader at BSC, said: ‘We do not have a good model to predict combustion and turbulence together with soot formation and evolution. We have challenges on the chemical, modelling and computational side because we need to compile a lot of different models while maintaining efficiency in HPC facilities.’

Lehmkuhl added: ‘We also have an extra challenge where, usually soot formation modelling is done with gas [engines] but, with kerosene, we have to take into account multi-phase conditions.’

There is also an industry-led requirement for this work, as Daniel Mira, propulsion technologies group leader at BSC, explained: ‘The aerospace industry has clear targets for CO2 emissions but now the ones for soot emissions are becoming increasingly important. The prediction of particle size distribution is a major challenge in the aeroengine industry, where standard prediction tools have massive errors between 200 and 300 per cent. We would like to see how these advanced soot models can predict soot formation at aeroengine-relevant conditions’

The team is making good progress but there are also many experimental considerations to take into account, where the team will eventually need to validate the model with experiments that successfully mimic real-world conditions. Lehmkuhl explained: ‘We must be able to correctly measure the soot formation and distribution in both laminar and turbulent flames, and this is very complex.’

The validation and assessment of the models will come from the definition of dedicated experiments with an increased level of complexity, going from pre-vaporised counterflow flames, kerosene spray flames up to single sector and full annular rigs. 

The project includes the development of reaction mechanisms for the fuel oxidation and soot precursors, and the development of a phenomenological model for primary breakup of airblast atomisers. ‘We have a lot of different physics that we need to put together, and we need to validate as much as is possible,’ Lehmkuhl added.

AI Potential

Could CFD benefit from an AI-enabled upgrade? Polish software specialist byteLake is developing algorithms to optimise existing CFD workflows and lower the cost of such solutions.

The byteLake CFD Suite features a collection of AI models, some replace traditional numerical solvers with equivalent AI models, while others accelerate CFD solvers with AI algorithms. This software suite promises to expedite the analysis process dramatically, providing ‘immediate results’, according to Marcin Rojek, co-founder at byteLake, who added: ‘Artificial intelligence is the next step for CFD, unlocking fast simulations.’

The firm is currently developing its CFD suite using steady-state simulations, before moving to transient simulations. ‘For steady-state, we have not seen any limitations where AI cannot be used,’ Rojek added.

The company recently completed its first accuracy tests with 24 different simulations using OpenFoam. A steady-state icoFoam simulation was carried out with a mesh size of 400 cells. The predicted results achieved a mean squared error of less than 0.001 and a Pearson correlation coefficient of more than 0.95, with more than 99 per cent relative accuracy. byteLAKE is now investigating larger mesh sizes and accelerating other OpenFoam solvers. 

Mariusz Kolanko, co-founder at byteLake, said: ‘The key challenge is to accelerate CFD simulations beyond what is currently possible. We are also working to ensure the AI models are easy to use and compatible with existing workflows.’

The CFD Suite is compatible with OpenFoam and uses TensorFlow-compatible algorithms to boost its interoperability, where TensorFlow is a leading open source machine learning platform. byteLAKE is working towards providing a plug-and-play interface, where CFD engineers should not need to change data types, replace toolchains, etc, while providing cross-platform compatibility, with no hardware or infrastructure upgrades required to run the software. 

Ease-of-use is another important deliverable for the CFD Suite, where engineers should just prepare their input data in the same way that they would for a traditional CFD solver but, instead, send it across to an AI-enhanced CFD model.

‘We want to make a tool that will be compatible with all the major CFD players and is also easy to use,’ Rojek said. The CFD Suite is scheduled for release in November 2020.

While challenges remain for today’s CFD solutions in terms of usability, validation and performance for a given accuracy, the inclusion of AI technologies is an interesting, new direction. This specialist work and other innovations are now more important than ever to address an increasing range of real-world problems in the CFD space. 

Exclude from view: