The smartest place I know: ‘This will be the decade of the digital twin’

What’s so smart about a digital twin? Perhaps the better question is, what hasn’t been smart about design and construction in the past? And the sad answer is: a lot. Over the last 50 years, society has achieved enormous productivity gains in almost every sector, but our industry has lagged behind.

With building information modelling (BIM), there has been a tendency to have different models. A digital twin unifies information in one validated environment that everybody shares. The idea is that there’s a golden thread all the way from the initial design through the selection of an item, its specification, its manufacture, its delivery to site, its installation and its whole life operation.

BIM was introduced 15 years ago, but increasing tech adoption in the construction industry has been accompanied by a decline in productivity. We didn’t remove inefficiency in the design process, we just digitized it. Digital twins hold the potential to actually change workflows. It’s not about replacing people, it’s about enabling them to do more complex tasks faster and better. By enabling us to really understand a design and optimize it, digital twins will reduce uncertainty, delay and mistakes.

The advantage of lagging behind is that you can learn from others. Back in the 1990s, the semiconductor industry hit a wall when computer chips got so complex that humans couldn’t optimize them any more. Now they do it by setting parameters for processing capability, memory and specialist functions and using the design engine to test many different configurations. A design team could never generate that many options. But now they can use their expertise to evaluate them, rather than spending time on the calculations.

Machine learning and AI will be important as this evolves, but we have to be realistic about what’s achievable. One limitation is our ability to define a good outcome: the more complex a design, the more sophisticated the parameters need to be. Then we have to feed those models — you need a large quantity of information and the ability to validate it to really train and hone those tools. With a skyscraper or a very large development like an airport, it would be very hard to do that, simply in terms of having enough comparable scenarios.

This will be the decade of the digital twin. There will still be outliers — the smallest bespoke scenarios that don’t support a twin, up to the largest, most complex scenarios. But even there, at the eight-to-ten year mark, we might see that people have more thoroughly adopted twins because they’ve become part of the workflow. We think that’s the biggest challenge: finding a new way of working that everyone can adopt, because that’s what’s going to drive change rather than the capability.

We’ve been reducing our carbon impact by eliminating concrete and moving to cross-laminated timber (CLT). The digital twin enables this kit-of-parts approach, but it works the other way too — it’s the modular and the more readily configurable that allows you to leverage what a twin can do. When you have to bring the right pieces to site at the right time and put them in the right places, you really need that golden thread. You also realize how much more efficient a team can be when they can say, “OK, today we know we’ve got pieces 21 through 46 arriving, and they go together in this sequence”.

The capacity to develop faster will be important for as long as we can foresee. Every day, a million people around the world move to cities. That creates huge demand in terms of where they live, work, rest. Where will they be educated and cared for? Where will the infrastructure come from? There’s an existential need for humanity to enable that urbanization to progress well. Digital twins are an exciting innovation, but they’re really just a necessity for responding to that demand.

Pete Swanson is head of strategic partnerships at Lendlease Technology