Last Thursday (19 March) marked the start of a new movement in shipping. For those of us that have an interest in ship model testing, the day when the experimenting carriage on the first towing tank to be built in the UK for over 50 years moved under its own power was a big moment.
It didn’t move far and it didn’t move fast, but it marked a significant step towards the facility finally being fully open. As far back as June 2015, it was reported in the media as being “scheduled to come into operation in the very near future”. After a number of delays it could be fully open a few weeks after the coronavirus crisis, its manager, Bertrand Malas, told me the following day. Just final drive system tuning and testing are still to be completed.
The tank is at the University of Southampton on the Boldrewood campus that it shares with Lloyd’s Register, although LR has not been involved in its development. Among other things, it will support students on its Ship Science course – of which I am a graduate – along with research and commercial contracts.
The previous British towing tank to open was on the Isle of Wight – just across the Solent from Southampton. It was built for the British Hovercraft Corporation in 1969 and was finally owned by GKN Aerospace, which closed it in 2008 to make more space for its aeronautical work. Some parts of it live on, however; its soleplates are doing the same job supporting the rails along the Boldrewood tank.
GKN’s tank was not the only one to close during the past half-century. Another was in St Albans, just north of London and owned by Vickers Shipbuilders. I had worked there for a couple of years after graduating. Other casualties included the three tanks operated by the National Physical Laboratory’s Ship Division, which later became the National Maritime Institute and was finally merged into BMT. One of its tanks was nearly 400m long and it was a visit to that establishment as a schoolboy in about 1973 that cemented my interest in ship design.
With so many tanks having closed and with computational fluid dynamics (CFD) becoming more sophisticated over the years, it is tempting to ask whether CFD has made physical modelling redundant. I raised that point during ShipInsight’s annual conference last month, which included a session on hull efficiency.
One of our panellists was Inno Gatin, CFD developer and consulting engineer at the fluid dynamics specialist Wikki, who said that although CFD had become more economical and practical in the past few years, it could not fully replace tank tests. They have become more economical, so CFD can do initial design work more quickly, but its results would “supplement the towing tank tests that come later in the design process.”
I put the same point to Mr Malas, even though I knew – from an excellent description of the tank’s development that he had written for February’s issue of the Royal Institution of Naval Architects’ (RINA) publication The Naval Architect – that this was hardly an original question for him.
But it is a particularly relevant question in this case because there are already ship model testing facilities a few miles away run by QinetiQ, formerly the Admiralty Experiment Works, where there is a large towing tank supported by an ocean basin and cavitation tunnel. So my question was twofold: why is a towing tank still relevant, and why here?
We took the second one first. Because of its university links, it will have three purposes: research, education and commercial work. “If one of these things hadn’t been there, the tank probably wouldn’t exist,” he said. The tank has already been used for all those purposes, even without a working carriage. In 2019 it clocked up 115.5 days of use, of which 81.5 were for education, 14 for research, 10.5 for commercial work and 9.5 for ‘open days and outreach’, according to figures Mr Malas included in a presentation on 5 March to the UK’s Southern Joint Branch of RINA and IMarEST. That Powerpoint includes video of the tank’s wavemaker in action and of some experimental work.
There are also practical difficulties with using QinetiQ’s tanks: it is about a 45-minute drive each way, which is too far for student work, and is too big. Availability can also be a problem: it is primarily used for naval work and can be block-booked for months at a time. Commercial customers face the same challenges, so “the fact that we can offer short term slots should be a big advantage for us,” he said. He predicted that the tank’s commercial work will grow to around 50-60 days per year over the next couple of years; “we will be a bit more flexible and hopefully a bit more efficient.”
As for my main point, he echoed the remarks made by Mr Gatin. “There are things that we don’t do in a tank any more because CFD is better and quicker,” he said. “But the vast majority of designs today still get a final set of validation tests in a tank.” For complex predictions, such as for seakeeping and manoeuvring tests, “I don’t believe CFD is good enough for that sort of thing yet.”
It seems, then, that a towing tank’s role has shifted in the years since I was in the business. Then, it was mainly to predict resistance and propulsion. Now, at the Boldrewood tank, “[what] we are working towards is to provide data to validate CFD” in addition to the more classic experiments, Mr Malas said.
The tank is not the only University of Southampton-based organisation offering ship performance predictions. The Wolfson Unit uses university facilities to offer model testing and consultancy on a commercial basis. It has been involved in developing the new tank – for example, by designing its sophisticated dynamometer and it will be involved in its commissioning – “but it’s not our facility,” its principal research engineer, Dr Sandy Wright, stressed to me.
The unit will probably be its biggest commercial user and Dr Wright confirmed Mr Malas’s view of its potential, saying “it will make us more competitive again.” At 138m, it is somewhat shorter than QinetiQ’s 270m tank, but that extra length has drawbacks: simply bringing the carriage back to the start takes more time than on the shorter tank, which means experimenting runs take place about every 25 mins, compared with an expected 7-8-minute interval in the Boldrewood tank.
CFD-only for America’s Cup
Dr Wright also has important things to say about the CFD vs towing tank debate. The Wolfson Unit is currently working for Ineos Team UK, which is developing the UK’s entry to next year’s America’s Cup competition. Its rules state (at paragraph 11.2) that “no competitor shall carry out fluid dynamic testing of designs, at any scale, in any controlled environment such as a wind tunnel, cavitation tunnel, towing tank or manoeuvring basin.”
That leaves CFD. Although this rule is included in a section on cost containment, Dr Wright suggested that it was drafted because New Zealand – which won the cup last time – does not have a suitable towing tank of its own. Mr Malas told me that New Zealand might think it has better CFD capabilities than others.
Dr Wright disputed that but conceded that New Zealand may have better simulation capability, which combines hydrodynamic, aerodynamic and crew data to produce a performance prediction. Whatever the reason, it means that the Wolfson Unit – housed in the same building as the university’s new tank – is contractually obliged not to use it for this project.
In fact, the Wolfson Unit’s task is to look at the above-water aerodynamics of the foil-supported craft, so the alternative to CFD in this case may have been wind tunnel tests – which it has used for previous America’s Cup entries – rather than a tank, but I asked him whether a CFD-only proposal could have won the America’s Cup contract if physical tests had been allowed. That depends, he said. If the project were entirely hydrodynamic, “it would be a brave person to run on purely CFD resources,” he said.
More generally, “CFD and tank testing can give you different things … and it will stay like that for a good while yet.” For example, CFD gives a better understanding of the physics involved in the flow round a hull and is satisfactory for modelling design variations, but “[if] you want some really solid numbers [for] a complex situation, then I would go tank testing.”
Tradition and standards
What CFD lacks, however, is tradition and standardisation. Tank testing has 150 years of history, Mr Malas pointed out, with procedures that are similar worldwide and backed by an association – the International Towing Tank Conference (ITTC) – that has observer status at IMO.
There are no such guidelines for CFD and “the quality of calculations you get is very variable,” he told me. He referred to his Naval Architect article in which he had mentioned a number of CFD workshops in recent years. At these, various CFD programs are given the same hull geometry to work with and their outputs are compared with results from model tests or sea trials. In his article, he noted scatter of up to 15% in the results, while ITTC guidelines say there should be no more than 2% uncertainty in model tests.
Those figures “are quite concerning,” he told me. Anyone can enter these workshops so perhaps they are not the highest quality programs, he suggested. If so, “where is the top quality? We don’t really know that.”
Work is being done to reduce this uncertainty, but CFD experts and tank testers “need to work together to produce the best overall results,” he said. At Boldrewood, this is possible, because both the tank’s experts and the Wolfson Unit are in the same building.
Dr Wright agreed, but had a slightly different perspective on those workshops. He acknowledged that CFD does not have the long history that tanks can boast, “but it is a lot better now.” The workshops might produce a range of results, but “if you average them all, it’s probably the correct answer, … whatever ‘the correct answer’ is.”
That is a good question: what is the correct answer? And how will we find it? This is where we came in. With towing tanks having played a formative role in my career, I am confident they will always have an important part in future ship technology.
It will be a different future from when I ran propulsion tests if tank tests simply become CFD confirmation tools for basic designs. But for the sophisticated designs that we may need in the drive to reduce shipping’s emissions, I doubt that anything will emerge to fully replace their exhilarating experimentation.
I hope I will get to experience the Boldrewood carriage one day: not at the slow pace of last week’s trial, but accelerating to its top speed of 12m/sec with its experimenters strapped to their seats. It will be quite a ride!
• What is your view of the CFD vs towing tank debate? Email me now with your views.