A hundred years ago, naval architects of the day had a simple preliminary test when designing a new yacht: if it looked right, it was right. Spool forward, and the face of naval architecture has changed almost beyond recognition as computer-aided design (CAD) and computational fluid dynamics (CFD) packages revolutionise the way hull forms are designed and optimised.
‘In my first year of university in 1995 we still learnt to draw with French curves, but computers were already in use,’ says Perry Van Oossanen, a naval architect who works within Van Oossanen & Associates, a design studio originally set up by his father who had been head of the advanced ships department at the Marin tank test facility. ‘In the 1990s big steps were made in their application to naval architecture.’
As one of the new generation of naval architects, Van Oossanen has embraced the application of technology to the science – and art – of yacht design.
‘A few years ago, you had to build a test tank model,’ he continues, ‘which could easily cost €10,000 each. But there is now a lot of confidence in CFD, and when we do build a tank model, we already know what to expect. In fact, my father says that if you see a big difference between a tank test and a CFD model, you should question the tank test results first…’
Van Oossanen has overseen the installation of a bank of computers at the studio’s offices in order to handle the complex calculations involved in CFD work.
My father says that if you see a big difference between a tank test and a CFD model, you should question the tank test results first…”
Perry Van Oossanen, naval architect, Van Oossanen & Associates
A more traditional tank-based study giving laminar flow visualisation along a hull form
A hundred years ago, naval architects of the day had a simple preliminary test when designing a new yacht: if it looked right, it was right. Spool forward, and the face of naval architecture has changed almost beyond recognition as computer-aided design (CAD) and computational fluid dynamics (CFD) packages revolutionise the way hull forms are designed and optimised.
‘In my first year of university in 1995 we still learnt to draw with French curves, but computers were already in use,’ says Perry Van Oossanen, a naval architect who works within Van Oossanen & Associates, a design studio originally set up by his father who had been head of the advanced ships department at the Marin tank test facility. ‘In the 1990s big steps were made in their application to naval architecture.’
As one of the new generation of naval architects, Van Oossanen has embraced the application of technology to the science – and art – of yacht design.
‘A few years ago, you had to build a test tank model,’ he continues, ‘which could easily cost €10,000 each. But there is now a lot of confidence in CFD, and when we do build a tank model, we already know what to expect. In fact, my father says that if you see a big difference between a tank test and a CFD model, you should question the tank test results first…’
Van Oossanen has overseen the installation of a bank of computers at the studio’s offices in order to handle the complex calculations involved in CFD work.
My father says that if you see a big difference between a tank test and a CFD model, you should question the tank test results first…”
Perry Van Oossanen, naval architect, Van Oossanen & Associates
‘We have 120 CPUs in our cluster, and it keeps on expanding,’ he says. ‘CFD is based on a 3D hull form model from our CAD program, and is a 3D computer model based on surfaces. The CAD model is exported to what is called a mesher program, which criticises the surfaces and translates them into millions of triangular cells, which we call a surface mesh.
‘Then the water is broken down into millions of cells, and another program called a solver works through the equations for each cell to create a picture of the fluid dynamics. Meshing takes from four hours for a bare hull to a couple of days for a complex hull with appendages, and we can do a hull test in a couple of hours.’
‘I started utilising the computer modelling program in the laboratories of Dassault in 1978 for my second America’s Cup campaign,’ says naval architect Philippe Briand. ‘At that time the data we obtained through computation was not very reliable. However, the tool proved to be outstanding for visualising the distribution of the zones of pressure on the hull. This allowed us to smooth out those pressures over the length of the hull.
‘Tank testing requires big models and to be accurate they have to be built to a minimum 1/8th scale. Over the years hydrodynamic scientific research and development has evolved to become much more precise and reliable. We are now able to create a mesh with 8 million cells covering the hull drawing. The precision of the resistance calculations has a deviation of less than 2 per cent compared with the results of tank testing.’
Of course, while CFD eliminates some of the problems of tank testing – the scale effects of waves, for example – it also has its limitations, and complicated sea keeping tests are still beyond the scope of existing CFD clusters. One can’t help but feel, though, that the days of the test tank are numbered.