Strain gauge...

< < Go back

Strain gauge...

Features

In the hands of experts finite-element-analysis (FEA) techniques can have a huge impact on the design process, optimizing structures in all sorts of ways. Very few people are capable of using it correctly, however, and current applications are perhaps more restricted than one would necessarily assume. Phil Draper quizzes High Modulus’s resident FEA expert James Anderson as to where we’re at with the technology and where it’s going.

AUGUST 2007

We’ve probably all seen some of the pretty pictures of structures that FEA (finite element analysis) software can produce. They look really neat and usually they’re oh-so-colourful. But, unless someone that really knows what they are doing has been responsible for generating them, they are likely at best to be worthless and at worst to be potentially lethal.
FEA work is complicated stuff. And thus far very little has permeated the marine industry at any level, despite what some people may tell you.
So what constitutes someone who knows what they’re doing with FEA? Is it someone who’s acquired some software and is confident enough to give it a go? Or perhaps someone who’s really done their homework and read the software manual thoroughly, maybe even twice? Maybe a graphic designer? An industrial designer? A yacht designer? No? Well a fully qualified naval architect, surely? No.
The only people that can really use FEA programs effectively are FEA specialists and don’t let anyone tell you otherwise. Yes there are various boatbuilders and yacht designers out there that have made, and continue to make, claims of having FEA capabilities, but, with all due respect to them if they should be reading this, they are merely spinning a line, jangling jargon.
What’s really scary are the ones that fall into any of the above categories and actually believe they know what they’re doing.
It matters not that they’ve actually managed to conjure up a wonderful 3D model showing a myriad of colours and some credible looking bright-red ‘hotspots’.
The only guys truly qualified to use an FEA system in anger are professional engineers, probably men or women with at least first degrees in a mechanical-related engineering discipline to their credit — and very probably one or more post-grad or professional qualifications besides — plus several years of ‘coal-face’ work under a suitably practiced mentor. If they’re outside the world of academia, they very probably will have had serious aeronautical or automotive stress work experience.
It will probably surprise most of those reading this that have actually heard of finite element analysis, and appreciate what some of those pretty pictures we mentioned earlier are meant to illustrate, just how little FEA has permeated this industry. Talk ‘cutting edge’ and most people in the boatbuilding world will immediately think America’s Cup. Whilst all the syndicates involved will have used FEA at some point during their design process, the amount of true optimization undertaken in comparison to that done, for example, on a Le Mans car, is tiny. This is partly due to the fact that the minimum thickness requirements define the majority of the skin thicknesses in the hull, though regions of differing fibre orientation will be used to maximize global bending stiffness. Although highly loaded regions will be refined and optimized as required, the lion’s share of the material in the boat is defined by the rule, rather than finely tuned using FEA. Even if the rules did not specify minimum thicknesses, it would be hard to use FEA to define hull laminates in ACC boats due to uncertainty in the loading conditions. How do the pressures vary along a hull driving through an unstable seaway? What is the dynamic amplification when you slam off a wave? Indeed the fact that FEA techniques have not been involved in the design of hull and deck structures very probably suggests why a few AC boats have broken in half over years, as pushing the boundaries has been a case of trial and error rather than an appliance of science.
But we’ll come back to such problems a little later.
For those that don’t know, finite element analysis is about considering how a structure reacts to applied loads on a ‘micro’ level — which is where the ‘finite elements’ come in. It is about considering a structure in terms of a mesh and evaluating the strains at work in each of those elements — and for strain read ‘load’ and ‘deflection’. The higher the stiffness the more load carried through that part of the structure and the lower the deflection.
The element size can be varied. Essentially a large grid will give a relatively crude picture and fine grid a much more refined analysis. But the finer the grid, the more complex the analysis and the more powerful the system needs to be to cope.
Plenty of areas can be over-designed, but it only takes one weak spot for a failure to occur.
So where to source FEA expertise?
Very few boatbuilders are big enough and or progressive enough to warrant their own dedicated FEA specialist — the Ferretti Group is one of the very few that springs to mind — so if you want access to it you’ll need to go the consultancy route.
Most of those with an interest in marine composites will have heard of High Modulus, for instance, which for much of its 28-year history has been one world’s leading consultancies specializing in composite structural engineering for the marine industry. So when we say that High Modulus, which in its offices in New Zealand and in the UK employs no fewer than 70 people, 35 of them professional engineers, has only had an in-house FEA capability for a little less than a year, it reinforces the fact that employment of such techniques in this industry is still extremely rare. And the decision to bring FEA in-house was not made lightly by High Modulus; only when they were confident they could do it properly, and had sufficient demand to justifying the investment in the highest quality tools and a resident FEA expert to undertake the projects.
High Modulus’s FEA capability stems from the recruitment of one man, Englishman James Anderson, whose expertise up until he joined the Auckland-headquartered firm had been honed in the aerospace and motorsports industries. Anderson was a graduate in aeronautical engineering from the University of Bristol, one of the UK’s finest higher education establishments. From there he spent a number of years on heavy engineering projects, most of which was related to failure analysis — trains in crashes, power stations in earthquakes and so on — simulations requiring non-linear calculations and dynamic testing. Next he went to a specialist FEA consultancy based in Sussex, England, that did most of its work for the motorsport industry.
Previously High Modulus’s FEA projects were subcontracted to Auckland-based Matrix Applied Computing, certainly the most revered FEA specialist in New Zealand.
“When I arrived the first FEA/global bending project done in-house at High Modulus took about a week,” he recalls. “It would have taken something like five weeks to get that one done outside… This was due to a combination of quicker communications and our ability to make judgment calls on which are the critical areas to be modelled in detail and which areas can be simplified. Everything is so much more immediate now and everyone in the team’s starting to appreciate what can be achieved and what the opportunities are…
FEA can run in parallel with a design project or it can actually drive the project.
“There are a lot of people out there that claim to have an FEA capability, says James Anderson. “And yes various CAD systems provide FEA ‘bolt-ons’… But being able to use a CAD system and being able to use a FEA tool are two very different things… If you put the wrong loads in, you can still get a pretty picture, but the usefulness of what comes out will be at best misleading and at worst potentially very dangerous… I wouldn’t let anyone that couldn’t do the necessary hand calcs loose with FEA… You really need to be able to get to within 20 per cent of the solution using conventional methods before you’re able to appreciate the sense of an FEA solution. And by sense I mean whether the answer’s sensible or silly…"
High Modulus makes use of Altair Engineering’s HyperWorks, an FEA bundle. This extremely powerful product, which consists of a variety of different tools, has with its origins in the automotive sector.
“We’ll take 3D geometry from designers and then convert that via HyperWorks to an appropriate 3D FEA model, which doesn’t have to be as detailed as you may think. These things work most efficiently with a cleaned up and simplified structure… Simplifying a structure for FEA requires an experienced eye not necessarily masses of time. It is about what to include and what to ignore… CAD geometry is about looking good. But FEA modeling is about making sure everything is properly connected. CAD systems can leave ‘cracks’ between surfaces and components. Cracks are no good for FEA purposes as they will immediately open up once loads are applied.”
FEA optimization techniques are extremely useful for concept evaluation, doing things quickly in the design process, in order to make things better, simpler and so on further down the road. “It is not a lumbering process at all,” says James Anderson. “Optimization can be extremely quick, if it is done in the right way and the right level of detail… Usually a good working solution means around 80 per cent of the perfect solution can be achieved in about 10 per cent of the time that it would take to do the same work using manual iterations.
As already mentioned, FEA works best if it is involved in the design process from the earliest stage. “But real life is seldom ideal,” admits Anderson. “Alas a more common scenario would be for a builder to say ‘I’ve bought the material we think we need, but we’re unsure what core we need or what to do with the inner skin thickness… We have to cope with such compromises a lot.”
Obviously FEA can a very useful tool for enhancing performance. But what is performance? Say, in the case of a sailboat, it could be sailing performance, price performance or ‘real performance’, which is almost always some degree of compromise?
Certainly classification societies have been slow to appreciate such advances. Hull and deck requirements are based upon experience. They tend to be defined using very conservative build codes from the likes of Lloyds, ABS, DNV, GL and so on. Thus far hull laminates are still being defined by simply panel theory.
Any FEA system will only be as good as the material data it is given. But getting real load properties from composite structures is much more complex than for metallic structures. What is nominally the same laminate could have a different fibre architecture, weave, resin system, cure cycle and manufacturer, all of which affect the final product. This why High Modulus has an extensive R&D department and from which is has built up an impressive database of test results over the years.
A recent High Modulus FEA project concerned the composite swing-over anchor-launching cradle for the new Ed Dubois-designed 44m (144ft) fast-cruising sloop Salperton, which was launched from Fitzroy Yachts in New Zealand in March. Keeping weight to a minimum in yachts such as this, particularly at the ends, is important. So goes the argument for that style of combined launcher and fairlead system and why it is being produced in composites. But probably the most important reason for the swing-over anchor is aesthetics. An open-section structure, where the direction of loads moves with the swing of the anchor chain, it was designed using carbon prepregs.
“The FEA job for us was all about minimizing the stresses in multi-load cases. With a part like that only around 10 per cent of the component is actually working structurally under any given load case,” says Anderson. “The rest of it is really there for practical considerations, such as toughness against knocks, and styling… That is why to all intents and purposes the part would look much the same regardless of the material choice… The finished black (carbon) structure weighed just 70kg,” says Anderson. “Had it been fabricated in stainless steel, it would have been dimensionally much the same, but would have weighed something over 300kg. If it hadn’t had been FEA optimized it would have probably weighed nearer 120kg… Less weight not only means lighter, but also less material and ultimately less labour required to lay the material in the mould.”
Modal performance analysis is another area where FEA can make a major contribution. For modal read vibration. FEA techniques enable dynamic simulations of structures, to evaluate the natural frequencies of a structure and move them away from those known to emanate from, say, main engines or generators.
Indeed one of the biggest FEA projects that High Modulus has been involved with thus far concerned the mast of an ultra-large motoryacht. Such is the size of that vessel that its top mast alone is some 14m (46ft) tall. Comfort through minimizing noise and vibration is essential on this sort of luxury yacht. That project represented three to four months work.
As to the future, we can all expect to hear a lot ore about FEA as software becomes cheaper and more widely available. More and more people will be using the tools, but this in itself will not lead to better boats. On the contrary it may lead some away from the safety of hand calculations and codes into the unknown where they are likely to make mistakes. Better boats start with a better understanding of structures and the loads working through them. In the right hands FEA is a powerful tool to help with the optimisation of marine structures.

For more, www.high-modulus.com.