IN FOCUS: Big data looks best in 3D

London
Source:
This story is sourced from Flight International
Subscribe today »

The death earlier this month of one of the more colourful characters in information technology marks a good time to consider one of the biggest challenges facing industry and government today, and one well known to aerospace decision-makers: too much data.

Hank Asher, who died aged 61, is credited by TLO, the Florida company he most recently founded, as the "father of data fusion", and as the inventor of online investigative systems that have helped law enforcement agencies in many countries sift vast quantities of data in often unconnected pools to find leads for otherwise-stymied detectives, including those who identified additional terrorists after the 11 September attacks and identified and arrested the so-called Beltway sniper, who terrorised Washington DC with a shooting spree in 2002 that left 10 people dead and several wounded.

JOINING THE DOTS

Twenty years before 11 September Asher had been smuggling cocaine in his private airplane in South America, before going on to work with the US Drug Enforcement Agency to convince Americans to get out of the trade. A long liaison with law enforcement ensued, which saw Asher the software entrepreneur providing the computing power behind some major investigations. As he has been quoted: "You could accidentally live next door to [11 September mastermind] Mohammed Atta. You couldn't accidentally live next door to Mohammed Atta twice."

dassault systemes 3d data view dassault systemes

Dassault Systèmes

If I get stuck in here, how much will it cost to design me out?

Sifting names, addresses, flightplans, etc to make connections between people is just one aspect of the challenge typically referred to as "big data". Indeed, Asher's company, TLO, describes itself as "a data solutions provider specializing in custom, scalable investigative and risk management tools for due diligence, threat assessment, identity authentication, fraud prevention and detection, legislative compliance and debt recovery".

For airframers, however, a different set of challenges prevails. Historically, they have embraced the advent of computer-aided design for the obvious reasons of speed, accuracy and the fact that, especially in the internet age, it allows engineers who are not in the same location to work on the same set of plans. The addition of three-dimensional visualisation improved the process - for example by letting designers see, at the initial design stage, whether or not a technician would actually be able to access key components through a maintenance hatch. By the mid-1990s 3D had advanced far enough to allow Boeing, with long-standing design systems partner Dassault Systèmes, to make a dramatic breakthrough by producing the 777 from a digital mock-up, doing away with the costly and time-consuming physical model.

As the 787 went into production in late 2011, Dassault Systèmes chief executive Bernard Charles remarked that modern 3D digital design software had advanced so far as to transform our concept of reality, musing that if Chinese counterfeiters were to buy a 787 and attempt to copy it, they would never succeed - but if they got hold of the digital plans, they could do it. One Boeing programme executive later told Flight International that remark missed the point, as while it would be miles better to try to counterfeit a 787 working from digital plans rather than measuring all the parts with a pair of callipers, the machine's dimensions are only one aspect of its true essence, which arguably resides in its 18 million lines of computer code as much as its shape and size.

Both points are valid, but Charles and Dassault Systèmes are thinking far, far beyond mere 3D design. By 1999 the 3D toolbox had introduced the concept of product life-cycle management, so designers could consider their creations' eventual scrappage. But in late 2012 these design tools were brought firmly into the big data era, with Dassault's launch of what it calls the 3D experience.

One simple example Dassault likes is coffee. Beans return a profit of a penny or so per kilo, but by the time they have been processed into a customer experience, Starbucks et al are earning many dollars per kilo. At one time, 3D software would have merely helped design an ergonomically successful coffee maker, but today it is also a tool that can help managers visualise complicated relationships between costs, pricing, customer preference, employee attitudes, retail location, supplier demands, etc - and, critically, understand the impact of changing any of these or other parameters.

RISK REDUCTION

In aerospace, the goal is to reduce the time, cost and risk associated with innovation. As Dassault vice president Pierre Marchadier puts it, airframers have to understand that airlines are not looking for "cool stuff - they are looking for efficiency". And, with 3D, he says, the airframer can show its airline customer the difference in time, cost and risk between various options when designing an aircraft. So, the customer's confidence in choices made early in a programme's life can be enhanced - and customer knowledge and preferences can be built into the programme.

But perhaps most critically, says Marchadier, 3D experience tools give engineers and product managers a way to visualise the impact, over the entire programme life, of choices they make early on. Engineers must always nail down some parameters early for design to progress, and leave others to be finalised later; the newest 3D tools, he claims, let engineers see the result in time, cost and programme risk of fixing any particular part of the design - to choose the ones that aren't hugely expensive, in money and time, to adjust later.

In software terms, the system extends the tools needed to analyse non-linear events - such as the failure of a carbon composite structure - to such problems as risk assessment or shifting customer preferences. That architecture, says Marchadier, allows the system to "embark all the key players", ranging from engineers and salespeople to airline passengers.

That's a lot of data, and it would be ambitious to imagine that managers and engineers will have the power to harness it completely any time soon. As Microsoft futurologist Dave Coplin recently told a Financial Times panel discussion on the challenge of big data, a lot of the data we have is of poor quality - but as we work with that data it will be improved, as will the tools and skills we employ to put it to use. And, he added: "All of a sudden it's going to come together."

Or, as Robert Madelin, the European Commission's director general for communication networks, content and technology, told a gathering of Dassault Systèmes customers in Brussels: "We can do things we can't imagine and because we can't imagine them we're not doing them yet."