When you think of tunnelling, you’re likely to conjure up images of giant tunnel-boring machines, metres in diameter and with names like Florence and Cecilia, slowly munching their way through the ground.
But visitors to the British Tunnelling Society’s conference and exhibition in London last autumn were presented with a very different vision of what tunnelling could look like in the future. Hypertunnel, a Basingstoke-based outfit launched in 2018, unveiled a radical new idea that proposes building tunnels using a combination of horizontal directional drilling, artificial intelligence and 3D printing.
The idea behind Hypertunnel came not from tunnelling experts, or indeed from anybody involved in civil engineering, but from a couple of businessmen whose background is – most recently, at least – in renewable energy.
Neither of the co-founders, Jeremy Hammond and Steve Jordan, has an engineering qualification, but their interest in renewable energy led them to conclude that many of the obstacles to making tidal-range energy work on a large scale in the UK could be overcome using a network of tunnels driven inland.
The problem, of course, was the massive cost of tunnelling. HS2 hasn’t revealed exactly how much Florenceand Cecilia – the two 170m-long TBMs currently boring 10m-diameter tunnels through the Chilterns – cost, but its own Guide to Tunnelling Costs, using 2011 figures and published prior to work commencing, quotes a price of between £16m and £18m each, plus additional fixed costs of between £35m and £45m.
Costs of this magnitude made Jordan and Hammond’s idea unviable. Jordan therefore set about looking for ways to take the cost out of tunnelling.
“I was amazed to discover that tunnelling methods had hardly changed in over a century,” says Jordan. With his layman’s vision unhindered by preconceptions, he looked at other industries where, he says, people were “handling things differently and doing elements of tunnelling faster”.
He explains: “I came at things from a different angle. For example, tunnellers see removing spoil from a tunnel as a problem, but if you are an ore miner, that is your revenue stream, so you get it out of the ground as quickly as possible. I asked, what can we learn from the way they do it?”
Jordan consequently went on to develop a concept that basically involves injecting the lining of a tunnel into the ground and then removing the waste. Work is carried out not just at the face, but throughout the entire length of a tunnel or underground space simultaneously, greatly reducing project timescales.
To achieve this, Jordan drew on techniques proven in other industries, including digital underground surveying, digital twinning, machine learning, 3D printing (also known as additive manufacturing), robotics and swarm techniques, all supported by artificial intelligence (AI) and virtual reality (VR).
Jordan says: “In order to maintain our fresh-thinking approach, it was critical from the outset that we didn’t employ anyone with a tunnelling background. We selected people with dynamic engineering experience in innovation-rich sectors such as Formula 1, aerospace, oil & gas and geoscience. We wanted to be limited only by our imagination; not constrained by convention.”
But to combine these disparate techniques and transform their concept into a viable reality, Jordan and Hammond knew they needed the expertise of qualified engineers and the support of the tunnelling profession.
Therefore they began presenting the idea to influential members of the tunnelling industry in a bid to sell the concept and acquire technical credibility. Surprisingly, perhaps, given the industry’s supposed conservatism, most of those to whom they presented the idea responded with enthusiasm.
If there’s one significant achievement Hypertunnel can already lay claim to, it is its success in winning the approval and support of some of the most highly respected figures in the engineering and infrastructure sector.
Among those recruited to Hypertunnel’s board of directors and its technical board are Mark Carne, former CEO of Network Rail, renowned civil and structural engineer Peter O’Riordan, known for his work on Crossrail 2, HS1 and HS2, and world-renowned tunnelling authority, Dr Kurt Zeidler.
By September 2019, Jordan and Hammond had assembled their technical board under the guidance of director of engineering Patrick Lane-Nott who arrived after almost 20 years in top-level motorsport where he worked with advanced racing car simulations, building digital twins for vehicle development and race strategy.
“The founders specifically didn’t want to involve any tunnelling people at this stage,” confirms Lane-Nott. “We spoke to a lot of tunnellers, but we didn’t employ any,” he adds.
Even so, when it came to drumming up support from the civil engineering establishment, Lane-Nott found himself pushing at an open door. “We haven’t had to do a hard sell on this. Change is often not well received in the construction industry, but it hasn’t been like that at all. The people we’re talking to are the type of people who can see the potential of the concept,” he says.
Although the whole Hypertunnel idea looks radically new, and talk of ‘bots’, ‘swarms’ and ‘digital twins’ sounds borderline sci-fi, none of the technology proposed by Hypertunnel is new or unproven. “We’re not reinventing the wheel here,” says Lane-Nott. “We’re just using the technology in a different way. Digital twinning is nothing new – I was using this technology 20 years ago in motorsport – but it’s become a bit of a buzz phrase at the moment.”
In fact, the phrase will probably be familiar to construction professional currently working on BIM-enabled projects. It essentially means building an interactive digital model of a real facility or structure – literally a digital facsimile – and then manipulating that to explore its limitations. “It allows you to change something in the virtual world without having to build it for real,” says Lane-Nott.
Many construction people might also recognise another crucial element of the Hypertunnel concept, horizontal directional drilling, though it will be far more familiar to people in the oil and gas industry. In Hypertunnel, the technology is used to establish the profile of the tunnel by drilling a series of parallel holes for its entire length. But after that, things rapidly get a bit more cutting-edge, as Lane-Nott explains:
Rather like the software bots that carry out automated tasks on the internet, these bots are small autonomous robotic vehicles that travel up and down the pre-drilled bot-ways performing a number of different tasks.
Some are ‘tractor-bots’, used to deliver materials and components to specific locations; others are surveying bots, used to investigate the geology surrounding the tunnel. And some are grouting bots, designed to drill into the ground surrounding their bot-way and inject grout to create a structural ‘shell’ that ultimately forms the tunnel lining.
Once this shell has been formed, the soil within is disrupted and removed with a remote-controlled excavator. A continuous concrete layer can then be sprayed onto the shell and the construction completed, if required, with custom linings. Secondary bore pipes can be used to house monitoring technologies to improve long-term tunnel maintenance and safety.
“Unlike traditional tunnelling, where you start at one end and keep digging until you reach the other end, we are working the whole length of the tunnel at the same time,” says Lane-Nott.
The bots are semi-cylindrical devices that can pass each other inside the bot-ways so that they can be deployed when required and never encounter a tail-back inside the bot-way. As Lane-Nott says, potentially thousands of bots could operate simultaneously inside the same bot-way.
And while to some of us this might sound implausible, yet again Lane-Nott insists that the technology is not new. “There are lots of bots designed to go down pipes already in other industries; though most of them are tethered while ours are not.
“However, our bots are not over-sophisticated. They are rugged, low-cost tools and as simple as possible. They only have to work up and down the pipe; they just go back and forth - they don’t have to navigate,” he says.
Nevertheless, they do communicate with each other. Using swarm technology, Hypertunnel can make all the bots share information and act in a coordinated way. “There is always some central control, but we want to make the bots as autonomous as possible. They will have some self-awareness; we give the bots a job-list and they go off and do it,” explains Lane-Nott.
Hypertunnel is still very much in the concept stage, though on-site trials have already been conducted and there are plans for a full-scale trial project later this year (see box, p47). But if it proves its viability, Hypertunnel could be used for a much wider range of applications than any traditional tunnelling method.
Lane-Nott says it will make underground construction easier, safer, quicker, less expensive and more sustainable as well as making existing tunnels easier to enlarge, strengthen, repair and maintain. “Comparative studies are difficult because this is so radically different. But we think Hypertunnel will be significantly faster and cheaper than current methods – and a lot safer,” he says.