Back to basics

Steve Howell, Abercus

October 1, 2016

FEA and CFD don’t have to break the bank, especially with open source tools now entering the market. Abercus’ Steve Howell explains.

Diesel exhaust plume dispersion. Images from Abercus.

It is increasingly necessary for companies to collaborate and innovate to reduce capital and operating expenditures, especially in this current lower oil price environment. Now is the time to invest in research and development to deliver improved performance and reduced costs. As part of this drive, advanced engineering simulation approaches including computational fluid dynamics (CFD) and finite element analysis (FEA) are being used increasingly within our industry.

By using first principles approaches, they can provide valuable insight at the design stage, improve safety, enhance understanding of installation and operational issues, and demonstrate technology readiness for novel products and approaches.

There are several commercially available general purpose CFD and FEA codes, developed over the decades, which have evolved into extremely impressive software tools. They are capable of simulating a wide range of complex physics and can be applied across many different industries, including the nuclear energy, aerospace, automotive, elite sports, built environment and offshore sectors.

These codes continue to be improved, requiring significant investment, ultimately funded by users through licenses. For some, however, these fees, and the associated investment required for the training of the engineers who use the simulation codes, might be a potential barrier to the adoption of simulation technology within their business.

Open source

Explosion analysis.

One trend that may assist with the wider adoption of engineering simulation is the emergence of lower cost and open source simulation tools. These tools are freely accessible to everyone and for many applications within our industry they offer a fit-for-purpose solution.

While the commercial general purpose codes are undoubtedly powerful simulation solutions, they may be unnecessarily gold-plated for many applications within our industry. If a company operates within the subsea sector, for example, and is interested in, say, wave loading on a subsea structure, then clearly it doesn’t necessarily need a CFD code that can simulate combustion and radiation, too.

Similarly, if a company is primarily interested in topsides technical safety applications, such as atmospheric exhaust dispersion or simulating fires and explosions, it may not need a CFD code that can also simulate the sloshing dynamics of a free surface within a separator, or the impact of a sand particle on a pipe wall and the associated rate of material erosion due to the impact.

For some companies, the use of open source software may provide a fit-for-purpose solution, but without the associated overhead of any software license fees. It is Abercus’ expectation that open source simulation tools will become increasingly used in the future, and that this will accelerate the democratization of advanced engineering simulation methods and their use in the offshore sector.

Gaining confidence

Verification and validation are the processes we must employ to gain confidence in our simulation models, to ensure that they are useful and fit-for-purpose. Verification is the process of determining that a computational model accurately represents the underlying mathematical model and its solution, whereas validation is the process of determining the degree to which a model is an accurate representation of the real world from the perspective of the intended uses of the model.

Verification is the domain of mathematics, demonstrating that the equations are solved correctly, whereas validation is the domain of physics, concerned with ensuring that the correct equations are solved for the application of interest.

By using a commercial simulation code, the code verification should already have been addressed because the software vendor should have delivered a code that correctly solves some set of documented equations. However, it remains the responsibility of the user to verify the calculation, for example, by demonstrating that the spatial and temporal discretization underlying the simulation is sufficiently resolved. When using an open source code, the user is responsible for verifying both the code and the calculation.

Regardless of whether the simulation tool is commercial or open source, it remains for the user to undertake validation activities to demonstrate that the equations being solved by the simulation code are actually the correct equations for the application of interest.

Ultimately, this requires a predictive capability assessment, which compares simulation predictions against some benchmark data from an experiment or some other engineering approach. Only if there is good agreement can there be confidence in the simulation approach for that particular application.

NAFEMS and ERCOFTAC

FEA and CFD in action. 

Anyone practicing in the field of engineering simulation should be aware of NAFEMS and ERCOFTAC. NAFEMS is the international association for the engineering modeling, analysis and simulation community, which covers the application of both CFD and FEA, and ERCOFTAC is a global association of research, education and industry groups focusing on the technology of flow, turbulence and combustion.

Both promote simulation best practice through the delivery of teaching courses and the organization of national/international conferences to promote the exchange of ideas and disseminate information. Since 1989, NAFEMS has published and maintained a quality system supplement to ISO 9001 which interprets the requirements of the ISO standard in the particular context of engineering simulation. In recent years, NAFEMS launched its professional simulation engineer scheme, which is designed to provide a consistent framework to demonstrate the competence of simulation users.

Both organizations provide repositories of benchmark data for the purpose of validation. Benchmark data is incredibly important for the purpose of validation activities and there is always a need for more reliable benchmark data. In recognition of this, NAFEMS and ERCOFTAC have recently established focus groups specifically for the oil and gas industry and associated simulation activities.

While there is often experimental testing undertaken during the course of projects, this tends to remain within the project. From the point of view of validation and general confidence in simulation this is unfortunate. Of course, for some applications it is necessary to keep experimental data guarded in order to protect the technology of interest, but, where credible experimental data is generic in nature, it could be shared with the simulation community, perhaps through organizations like NAFEMS and ERCOFTAC. This could help to significantly advance confidence in CFD and FEA across our industry.

Conclusion

CFD and FEA are first principles, advanced engineering simulation approaches that have much to offer the offshore sector. Traditionally they have, perhaps, been regarded as high cost niche simulation tools. But, it is expected that the emerging open source tools will help to accelerate the democratization of these methods so that they become more mainstream within our industry. With the increasing use of engineering simulation, it is crucial that as an industry we are rigorous with respect to verification and validation in order to maintain confidence in the simulation approaches and tools that we use, and in the predictions they yield.


Steve Howell
is technical director at Abercus, based in Aberdeen. He has held senior roles at Prospect Flow Solutions, Cundall, Mott MacDonald and Mobius Dynamics. He has a PhD in computational fluid dynamics from the University of Newcastle-upon-Tyne.