The idea of design for manufacture is hardly new, so why is manufacturing so often an afterthought for designers, asks E&T.
Fifty per cent funded by the Technology Strategy Board and nine regional development agencies, an ambitious project known as the Next Generation Composite Wing seeks to keep the UK at the cutting edge of innovation in aircraft wing development.
And not just any aircraft wing. The market that the project has in its sights is the single-aisle replacement fleet - in other words, aircraft to replace Boeing's best-selling 737 and Airbus' equally ubiquitous A320 family.
The commercial possibilities are obvious, as is the potential for the British economy. Each model is produced in volumes approaching 40 aircraft a month, with GKN Aerospace's wing component factory at Filton and Airbus' wing assembly plant at Broughton carrying out the bulk of the A320 wing manufacturing work.
And so, led by Airbus, the £103m three-year Next Generation Composite Wing project harnesses 17 leading British organisations - aerospace companies, consultancies, research bodies and technology firms - to a single goal: revolutionise aircraft wing development.
In part, of course, a lot of that revolution will be in the form of increased aircraft performance. Built of composites, wings will be lighter, as well as stronger. In short, aircraft will fly further, faster and more fuel-efficiently.
But that's not all. For, in stark contrast to how wing development has taken place in the past, a far from insignificant proportion of the project's output will be generated before a single aircraft takes to the skies. One of those research partners - Greater Manchester-based privately-held specialist aerospace firm Hyde Group - it turns out, is tasked with optimising the wing's manufacturability.
Employing advanced simulation software sourced from another of the 17 partners, aerospace giant Dassault's DELMIA simulation subsidiary, Hyde's role is to strip cost out of the manufacturing process - recurring costs, such as labour and materials, as well as non-recurring costs such as tools, jigs and fixtures.
In part, explains Hyde Group technical director Richard Waring, the task involves straightforward simulation: building a 'virtual factory', complete with virtual operators, virtual robots, virtual machines and virtual handling equipment, and then developing optimal manufacturing processes that are capable of producing at the required run-rate of 40 aircraft wing sets per month.
The level of detail is immense, he adds, and involves code development for the robotic cells and machinery that will carry out key parts of the manufacturing process, as well as lineside operator instructions and establishing the optimum ergonomics of loading parts into jigs.
But it's the two-way interaction with the wing's designers that is arguably the most exciting part of the project. For as well as optimising the production process required for a given wing design, explains Waring, Hyde's team also feed information back into the design process, exchanging 3D data to highlight how potential changes could reduce the build time, or take out assembly cost - or both.
The result is clear. 'Designers are better able to perceive and consider the constraints of manufacturing at the design phase, thus removing risk from later stages,' sums up Waring. 'Using simulation software in this way we can reduce overall costs by taking time out of processes and making better use of resources.'
Yet while hard numbers are difficult to come by, anecdotal evidence suggests that the sort of design-for-manufacture processes involved in the Next Generation Composite Wing project is very much the exception and not the rule.
As a concept, design-for-manufacture as a lean-inspired manufacturing improvement technique goes back to at least 1990 - and yet, 20 years on, seemingly little has changed in day to day design-for-manufacture practice. Faster set-up times, kanbans, zero defects and Six Sigma: these and other similar initiatives have undeniably wrought significant transformations on the manufacturing landscape. Yet - in large part - design-for-manufacture has not.
This is surprising because, like motherhood and apple pie, design-for-manufacture is one of those strategies that it seems difficult to argue against. In short, a product designed from the start to be easy to manufacture will be consequently cheaper to manufacture, quicker to manufacture and have better yield and quality characteristics.
'Design a product with an eye on its manufacturability, and there's the potential to save serious amounts of money,' says David Wolfe, a consultant in the manufacturing services group at Cambridge-based electronics design consultancy Plextek. 'If it's easy to build, it will be easy to test - and you'll get the yield up more quickly, too.'
But time and again, the evidence suggests that designing purely for function is still all too common - with the manufacturing, manufacturing engineering and procurement functions being involved late in the design process, if at all.
'There's still a significant disconnect between engineering and manufacturing,' agrees Paul Fewtrell, Manchester-based head of the Manufacturing Advisory Service North West and a former Toyota employee with significant design-for-manufacture experience. 'And the result is constant re-designs to meet manufacturing's requirements - re-designs that add cost and waste and that would be avoided if the design was right first time.'
It's an observation that cuts right to the heart of the debate around design-for-manufacture. Despite design-for-manufacture's compelling logic, designers still aren't talking to manufacturing. As Fewtrell notes, generic rules that address the requirements of EMC, fire safety or machinery directives are routinely incorporated into design activities - but not, strangely, design-for-manufacture.
But why does this continue to be the case? What practical steps can manufacturers take to remedy matters? What might an effective design-for-manufacture process actually involve? And do the new IT tools and production techniques available in 2010 make a difference - or is the battleground effectively the same as in 1990?
Talk to people close to the issue and it quickly emerges that the problem doesn't lie with the availability of design-for-manufacture tools and methodologies. In large part, what needs doing has been known for 20 years. And advances in IT - as with the Next Generation Composite Wing - make it easier than ever to share designs between engineering and production, and test and improve the manufacturability of those designs.
Nor is the affordability of those IT tools a constraint. Thanks to massive improvements in processing power, graphics processing unit performance, network bandwidth and solid-state drives, today's 64-bit machines are - literally - supercomputers on a desktop. CAD vendor Autodesk, for instance, estimates that a computer task that would take an hour on a standard desktop computer about eight years ago now takes six seconds.
Instead, the problem with design-for-manufacture seems to boil down to a combination of competing priorities and organisational issues.
'The norm - especially in large companies - is for each department down the chain to jealously guard its specialist viewpoint, often at the expense of the greater good,' asserts Gus Desbarats, chairman of TheAlloy, an industrial design consultancy with a client base that includes Hewlett-Packard, Toshiba, BT, Dixons and PZ Cussons.
'The result is 'knowledge silos' where the original design is created without any reference to manufacturability. Succeeding waves of technical specialists are then stuck doing damage limitation - causing delays, and then sometimes the destruction of the original proposition through 'accidental' and unintended changes.'
And competing priorities compound the problem. The two design 'missions' - design-for-manufacture and design-for-function - have very different goals and objectives, bringing using tensions and conflicts in their wake.
'There are things that matter to a design manager - and very different things that matter to a manufacturing manager, or to a procurement manager, who has to source the materials and components for a design,' says Tom Bianchi, European marketing manager for Dassault Systemes' SIMULIA design simulation software application. 'The agendas are set at a functional level, not an organisational level.'
The Manufacturing Advisory Service's Paul Fewtrell agrees, pointing to a combination of organisational issues and time pressures conspiring to keep designers and manufacturing people within their respective functional boundaries.
'There is often a certain arrogance on the part of the design team ('Who do these manufacturing guys think they are in telling us how to design my product'),' he says. 'This then gets coupled to an element of naivety on the part of manufacturing in not properly understanding the implications of their suggestions on how to make a product easier to assemble or test, for instance.'
Once past a certain stage in the process, he says, it becomes easier to resist improvement than incorporate it.
Interaction and communication
So what can be done to improve the prospects for design-for-manufacture? Talk to experts and several strategies emerge.
First, encourage interaction between design and manufacturing. The Manufacturing Advisory Service's Paul Fewtrell, for instance, advises getting people from the design team to spend a few hours on the production line. 'Not just once, but every six months or so,' he urges. 'It builds up a rapport and increases the level of understanding of the production process within the team.'
And second, if design-for-manufacture is still more of an aspiration than a reality, then try coercion. At Liverpool-based Brainboxes, a designer and manufacturer of specialist serial devices, for instance, it turns out that a highly formalised project management system enforces consideration of design-for-manufacture issues.
Critically, says technical director Eamonn Walsh, the project management methodology mandates a review of each design by production - and early enough in the design process for it to be meaningful. 'Production have to look not just at a 2D representation of the design, but an on-screen 3D representation, too,' he says.
And with a formal sign-off of the design as manufacturable, the logic is inexorable: production can hardly subsequently complain about a design's manufacturing characteristics when it's in production, if they previously approved the design as manufacturable.
Thirdly, start quantifying the cost of poor manufacturability. 'Resist the urge to correct existing designs, but use the strong economic facts to strengthen the case for the inclusion of design-for-manufacture tasks and reviews in the projects that you are planning now,' advises Fewtrell.
And finally, it's always possible to buy-in design-for-manufacture. As companies - especially in hi-tech areas - increasingly outsource their manufacturing, subcontract manufacturers are finding a ready market in making designs more manufacturable.
'Our customers know their markets and their customers, and know how to design products that meet the requirements of those markets and customers - but don't necessarily know what makes a product straightforward and efficient to manufacture and test,' says David Taylor, an electronics engineer and director of special business services at Southampton-headquartered contract electronics manufacturer ACW Technology.
'The earlier we can engage with them, with our engineers talking to their engineers, the easier it becomes to optimise designs for a manufacturing process.'
That said, sums up TheAlloy's Gus Desbarats, always remember that design-for-manufacture versus design-for-function involves a balancing act - which manufacturers ignore at their peril.
'Every conversation about design-for-manufacture needs to start with a clear understanding that if a product doesn't deliver a great experience for customers, no one will care how efficiently it is made,' he warns. In other words, never forget that designing a great product is every bit as important as designing a product that's easy to manufacture.