Printer-friendly versionPDF version


At any marine conference or workshop, you will hear many topics being debated: validation and accuracy of CFD predictions, best practices in simulation, the latest industry regulations affecting ship design and many more. One topic that always gets engineers talking and arguing is whether to run your CFD simulations at model scale or full ship scale. There are benefits and drawbacks to both approaches.

Running at model scale means the result can be easily validated by comparison with towing tank data before running new analyses. The model scale results can then be transformed to full-scale data using standard semi-empirical scaling up procedures. However, this approach can introduce uncertainties as the scaling methods may not be suitable for new or unusual geometries or running conditions. This is one of the reasons why ship designers are reluctant to change anything in their designs for which they have collected operational data and know the performance. Running CFD simulation at full scale may require a higher grid resolution to accurately capture aft end flow especially when appendages are present, but removes the possible errors from the semi-empirical scaling methods. It is however much harder to validate full-scale CFD results. One reason for this, as this video shows, is that collecting full-scale ship data is not an easy process!

The American Bureau of Shipping (ABS) is the latest company to share results of some tests they ran to help answer this model scale vs full-scale debate. They used STAR-CCM+ to run simulations on multiple hull variants, based on the KCS 3600 TEU container vessel geometry. ABS ran the same cases at both model scale and full scale, then compared the results, along with extrapolated model scale to full-scale results (using ITTC 1978). The results (shown below) gave two conclusions. For situations where viscosity is not dominant, such as bulbous bow investigations, all three methods ranked the different hull variants in the same order of efficiency (figure 1).  In cases where hull changes mostly affect the viscous part of the overall resistance, such as aftbody optimization, each of the three methods gave different results: as figure 2 shows, the ranking was completely different for each method (though the relative differences were small in most cases).


Figure 1: Ranking of the five best geometries for a bulbous bow simulation, run at Model Scale (MS), Extrapolated Full Scale (EFS) and Computed Full Scale (CFS). All three methods ranked the same five geometries in the same order.


Figure 2: Ranking of the best 5 geometries for a skeg simulation, run at Model Scale (MS), Extrapolated Full Scale (EFS) and Computed Full Scale (CFS). In this case, all three methods ranked the top geometries in a different order.

This latest finding has added to our understanding but does not give a definitive answer which of the results we should believe. It clearly depends on how much confidence we have in our CFD predictions. Lloyd’s Register’s recent full-scale hydrodynamics workshop showed a remarkable alignment of CFD results from multiple participants with full-scale sea trial data, showing that we have the tools within our industry to accurately and consistently model at full scale. To ensure consistent results, of course, requires rigorous best practices for model setup, another point covered by ABS in their recent webinar.

With this presentation from ABS, we have now hosted webinars from all three major classification societies on their use of STAR-CCM+. All the societies are showing clear leadership in their application of CFD, but it is interesting to note the different aspects they chose to present. Lloyd's Register’s focus was on full-scale simulation, DNV-GL has investigated the added resistance in waves, and ABS are emphasizing process automation and optimization. While most companies are still using CFD to validate and troubleshoot physical designs, ABS has shown they have taken their use a step further by automating their process: driving STAR-CCM+ via macros to automate the setup and running of many cases in one session. This approach means they are using CFD in an effective and efficient way to rapidly evaluate multiple ship hull geometries. This shifts CFD away from its traditional use analyzing and troubleshooting a single design, instead, showing its power as a design exploration tool. 

I am sure the debate on model scale vs full scale will continue for years to come. But what cannot be denied is the power of CFD to aid decision making through design exploration, when choosing the best performing hull. As the top three Marine class societies lead the way in validation and best practices in the use of CFD the rest of the industry is now rapidly adopting CFD methodology as well. This drive towards automation and design exploration, combined with more realistic operating conditions achieved through full-scale simulation will enable us to evaluate novel, advanced and efficient design solutions, leading to more fuel and cost savings as well as much cleaner operations in our industry.


Matthew Godo
STAR-CCM+ Product Manager
Stephen Ferguson
Marketing Director
Brigid Blaschak
Communications Specialist
James Clement
STAR-CCM+ Product Manager
Joel Davison
Lead Product Manager, STAR-CCM+
Dr Mesh
Meshing Guru
Ravindra Aglave
Director - Chemical Processing
Sabine Goodwin
Director, Product Marketing