Traditional ship design is based on testing downscaled models. At full scale, behaviour often differs from test results. Dejan Radosavljevic, CD-adapco’s Director Marine, emphasizes the role of Computational Fluid Dynamics (CFD)
Mr. Radosavljevic, when did CFD become interesting for the marine industry?
Dejan Radosavljevic: Back when I was still[ds_preview] working for Lloyd’s Register, we investigated failures that occurred during ship operation. In search for the root cause we very often found a direct link between design being done at model scale and not being able to identify the problems that could appear at full scale (scale effects due to violation of similarity law governing viscous flow, editor’s note). So very early on we started applying full scale CFD.
For the wider marine industry CFD became interesting a couple of years ago when the oil price went through the roof. Suddenly they realized that they had to design much more efficient ships and started looking for alternative methods. Up to that point, it did not matter so much how well you designed your ship as long it achieved a certain speed, the only contractual requirement. There was no pressure or incentive to make sure the design did not use a bigger engine or consumed more fuel than necessary.
Now the oil price has gone through the hole in the ground but environmental regulations and climate protection today cause an even bigger demand for efficient design. Traditional design solutions do not seem to offer enough gain in efficiency. This new technology really allows you to innovate as much as you want.
What position does CD-adapco have in the marine market today?
Radosavljevic: I believe that with STAR-CCM+ we are the technology leader in the marine market but in terms of penetration it is really hard to tell, because there are still many legacy codes or in-house codes in use in many companies. People find it difficult to move away from something that they have been doing for so many years. So you can say our biggest competitor really is tradition. But what I believe though, is that the drivers to improve design will finally affect everyone.
We have seen that industry leaders, e.g. Korean shipyards, have moved to CFD and are using our product STAR CCM+. At that level of advanced shipbuilding I am confident that we have by far the biggest share. In less advanced parts of the market we probably do not have that much penetration because of inertia and people sticking to methodology they have been using for many years.
What prevents more progressive design?
Radosavljevic: Let us look at air lubrication as an example: The idea existed for a long time but no one was prepared to take the risk to build it in full scale, measure the effects and find out whether it works or not. With a new idea you traditionally have to go to the towing tank, but nobody can tell you what your model will do in full scale as there are no previous trial’s data to help with scaling.
Another example is true multi-objective design optimisation where you want to test hundreds of different solutions seeking the best performing one across all your optimisation criteria. You cannot build a scale model for every design that looks promising in order to test it in the tank to check how it compares in terms of performance. CFD can do all that but there is still an issue of confidence in the results. Not as much between the designers and CFD but primarily between the shipowners and CFD. The future is full scale CFD and the owners really have to open up and start to trust and accept the value of it.
Are the physics of water and wind fully understood and translated into CFD software or is there still need for further research?
Radosavljevic: The physics is pretty much well understood. What is not fully understood is, for example, how to predict freak waves when they happen out in the ocean. But that is more applicable to offshore structures than the actual ship. Ship design strength is based on »significant wave height« and you can use CFD to describe any kind of wave they could encounter within this context. But bear in mind that these simulations are still time consuming, even if processors are getting better and better. The way to overcome this for now is to use some clever way to first identify critical scenarios and then run full simulations only for this selection.
Will there be a future of ship design without tow tanks and cavitation tunnels?
Radosavljevic: They will always have a place. But I think the emphasis will be more during education of naval architects and for aspects of fundamental research, for example it will always be useful for validation of new numerical methods, because a towing tank is the easiest way to have a controlled environment. And there are certain aspects of manoeuvring and wave encounter that at present can be tested faster in the tank. But I think in general the emphasis will be reversed from doing everything in the tank and a little in CFD to doing most of the work in CFD and just a little bit in the tank.
Felix Selzer
