Many will be interested in the question, how accurate are CFD simulation results? The answer: it depends. Whilst the results from manual or traditional engineering calculations are relatively easier to check, unfortunately, due to the complexity involved in the iterations of partial differential equations, checking CFD calculations can be tedious whereas conducting physical validations to determine the accuracy of CFD results may be impractical and/or economically prohibitive. Consequently, stakeholders may have to rely on CFD results per se to make informed decisions that may involve safety and/or financial consequences. Obviously, having accurate CFD results will inspire confidence in the decision-making process.
The accuracy of CFD results depends on six factors, namely:
- Boundary conditions employed
- Engineering knowledge of the subject matter
- Experience and track record in performing similar CFD projects
- Qualification and experience in performing CFD simulation
- Quality control system
- Simulation software employed
One of the most important factors that determines the accuracy of any CFD result is the boundary conditions (BCs). Each of the mathematical equations requires meaningful values at the boundaries of the fluid domain for the calculations to generate reliable results. These numerical values are known as the BCs and can be specified in several ways, although in general, the specification of multiphase phenomena or phenomena involving reactions is more complex than single-phase phenomena.
The use of wrong or inaccurate BCs will render the results inaccurate and must be prevented before modelling and simulation commence. Hence, it is highly important that the BCs employed are meaningful, and accurate, and will lead to results that meet the objectives of the CFD studies.
Every stage should be checked and reviewed to ensure the inputs or results are numerically accurate and make engineering-sense
Next, it is important that the engineer conducting CFD analysis has relevant knowledge of the subject matter to be able to appreciate the results obtained through CFD simulation. Having relevant knowledge of the subject matter will also enable the engineer to know what results to extract and to perform meaningful interpretation of the results.
A strong track record in delivering similar CFD projects supported by the engineer’s industry background is another indicator of the suitability of the engineer to perform a particular CFD analysis.
In contrast, performing CFD analyses without proper knowledge may lead to misleading results.
It is important that only engineers with relevant qualifications and experience in CFD are engaged to conduct CFD analyses.
To obtain accurate CFD results, the CFD analysis process must be subjected to stringent quality control system. Every stage should be checked and reviewed to ensure the inputs or results are numerically accurate and make engineering sense before being allowed to proceed to the next stage.
This strategy prevents small errors emanating from each stage to snowball into a large error which ultimately affects the accuracy of the final results. Upon completion of all modelling and simulation iterations, the final results should be independently reviewed by another engineer who has equivalent or more experience in conducting similar types of CFD analysis.
Finally, the CFD simulation software employed should have a strong track record of being accurately validated against independent physical experimental studies.
Only when all six factors are observed will the CFD results be accurate enough for stakeholders to rely upon them, without the need for additional validations, to make informed decisions.