Detecting Leaks from Outer Space08 April 2019 | Facts & Curiosities
A desert is the last place you’d expect to worry about water leaks, but in arid countries like Libya, a burst pipe can literally be a matter of life and death.
Beneath its barren plains and blistering sands, thousands of miles of pipework carry life-giving water from aquifers and treatment plants to remote, isolated communities, enabling life to continue amidst the unforgiving desert. A burst pipe can spell disaster for an entire community, but checking and maintaining the country’s water infrastructure is no easy task when summer temperatures top 50 degrees C and skirmishes between armed militia and criminal factions are rife.
If you’re a regular reader, you’ll be well aware that here at Hero Labs we’re huge believers in using technology to solve real-world problems – doubly so when it comes to anything to do with saving water (or money for that matter).
So we were intrigued to read about a ground-breaking partnership between researchers at the University of Nottingham and the government of Libya – using autonomous quadcopters, blimps and even Low-Earth-Orbit satellites to pinpoint leaks from the air. The “UAVs” – autonomous airborne vehicles, or drones to you and me – were equipped with infra-red cameras that enabled them to make thermal maps of vast tracts of land from the safety of the sky. Flowing water creates a cooling effect, so a burst pipe in a blazing desert appears as a great blue smudge on the image and hey presto, you have the location of your leak down to a few dozen square meters.
“Water is one of the most precious commodities in the world”
says Professor Amin Al-Habaibeh, from the School of Architecture at Nottingham Trent University, “but we lose vast quantities every day to leaks and burst pipes. This is a high-speed, non-destructive way to survey vast areas of desert at once”. Professor Al-Habaibeh’s team hopes that the technology can help to revolutionise leak detection in one of the driest countries in the world, improving life for pipeline workers and remote communities alike.
It turns out the University of Nottingham Project is not alone. South West Water (SWW) recently teamed up with the University of Exeter to launch a pilot scheme (or should it be pilotless scheme?) using unmanned drones with infra-red cameras to patrol its 18,000km of underground pipes for leaks.
The South West of England may not conjure up the same sense of peril and adventure as the blistering sands of the north-eastern Sahara, but it does have its fair share of rolling hills, country roads and remote rural communities, which come with their own set of challenges. SSW counts five of the ten least populated counties in England within its catchment areas, but also more than 25% of all of the country’s farmland – so demand is high, distances are large, customers are remote and there’s no room for downtime or the crops might fail. Bob Taylor, Director of Drinking Water, explains: “traditional methods aren’t always cost-effective out here – but water is precious. We’ve got to continually dream up new ways to detect leaks at a landscape level”. Since adopting technologies like drone-based surveying, SSW says they’ve cut leakage by around 40% – with performance sitting at more than double the industry average. Great work, guys!
But perhaps even more impressive is Severn Trent Water’s latest initiative – using a Japanese satellite orbiting 637km above Earth to spot escaping water from outside the atmosphere. Rather than using infra-red to detect temperature differences, it looks out for the chlorine in treated water – or rather the characteristic longwave radiation that chlorine emits. The satellite snaps pictures that cover around 3,500 square kilometres at a time and beams them back to a data science lab in Israel where clever algorithms and very powerful computers do their thing. The result is a map of “spectral signatures” showing where chlorinated water is present; overlay that onto a map of all the pipelines you own and play spot the difference to find your leak.
The resourceful company behind the technology can even use their eye-in-the-sky position to monitor for changes in vegetation that might give away a slow leak or one deep underground. If there’s a sudden burst of green growth in the middle of a desert, water must be coming from somewhere. Apparently, it works at night and irrespective of weather conditions too. Pretty awesome! It’ll be a while before you can use this to tackle your leaky taps, though – most customers get one set of images a year, and the maximum possible frequency is once every three months due to the satellite’s path over Earth. Doesn’t help you too much when there’s water gushing out of your ceiling (although there is an equally genius solution for that).
From blue skies to brown earth, some companies are looking lower for the solution to water leak woes – companies like the wonderfully named “Superdroid Robots”, who make remote-controlled robots that explore hard-to-reach pipe networks and transmit footage back to an operator on the surface. There’s also Inuktun’s “Versatrax 150”, a cute little feller that’s decidedly more Star Wars than Skynet in the looks department. He even bears more than passing resemblance to Disney’s Wall-E at first glance. The Versatrax can carry cameras, lights, gas monitors, drills, cutting tools, lasers, sharp metallic claws… okay, perhaps it’s a little bit Skynet. There’s even a Norwegian robotics outfit experimenting with snake-like, articulated drones that swim and slither around tight bends and through clogged pipes to overcome the limitations of tracks and wheels – and the need to arm your definitely-not-evil robot with sharp objects (we’re looking at you Inuktun).
Amazing as they are, most of these electronic explorers cost around half a million dollars, so don’t put them on the Christmas list just yet. But a Dyson-award-winning invention out of the Massachusetts Institute for Technology promises to be a game-changer in that respect – the tiny flexible robot is billed to cost just a “couple of thousand dollars”, can autonomously roam pipelines without a trained operator, and for bonus points, it looks rather like a brightly-coloured badminton birdie. The genius of its design lies in the fact it harnesses existing water flow to float effortlessly around the network, rather than relying on complex traction systems and hefty battery packs.
Another watery contender for this year’s Dyson awards was Infinite Cooling, a technology that recovers massive quantities of clean water from power plant cooling tower plumes. Power plants use massive amounts of fresh water – even our sci-fi-like nuclear power plants, which harness the very power of the atom, still work by heating water into steam to drive enormous turbines. In the US, 39% of all fresh water goes to power plants – and Infinite Cooling’s technology promises to reduce wastage by as much as 80%, so it’s hugely promising. “An average powerplant here in the US uses 13,000 litres of water a minute”, says founder Maher Damak, “and right now all of that just evaporates away. And at the same time, we’ve got droughts and wildfires across the country because aquifers are drying up”. Real-world trials start soon, and we’re excited to see what it can do.
If all this talk of water is making you thirsty, perhaps a leak detection technology that cut its teeth in Coca Cola’s bottling plants will quench your curiosity.
The system, dubbed “Leaf”, was originally developed by AI research lab I-Systems to control robots on production lines. It uses a branch of machine learning called “fuzzy logic” to observe changes in the operating environment and tweak variables accordingly, automatically managing the whole process and learning as it goes. It reduced leakage at the Coca Cola plant by about 2.5 million litres a year, and presumably made things considerably less sticky for its poor employees in the process – but I-Systems had bigger fish to fry. Armed with a database of “learning instances” from their time in the factory, they weened their AI off fizzy drinks and got to work on water.
In much the same way as it once optimised factory robots, Leaf can talk to the various valves and pumping stations that make up a modern water distribution network. The AI considers inputs like the time and day of the week, local temperature, cloud cover, whether it’s rained recently and even calendar events like national holidays and festivals. It then attempts to predict how much water will be used in any given area, and supply just enough pressure to meet demand. The lower the pressure, the less water that sprays out of faulty plumbing and leaky appliances – so leakage can be reduced across the whole network just by fine-tuning supply pressure to each region. If it’s a sunny summer holiday, people flock to the beach – so the coast might need more pressure, but it can probably be turned down inland. When it’s cold and raining, the whole network uses less water so pressure can be reduced throughout – saving electricity as well as water. The really cool part is it learns as it goes, so it just gets better over time. In one test in Brazil, it reduced leakage by 5.8% after just one week, and at points hit nearly 15%. That’s pretty smart!
If you like that, there’s another AI-powered assistant will probably float your boat – our very own Sonic, of course. Much like Leaf, Sonic uses advanced AI algorithms to learn what normal water use looks like for your household and figure out ways to help. He can spot all kinds of leaks ranging from burst pipes to pinhole leaks, and he’s even smart enough to recognise different appliances like dishwashers and washing machines and estimate how much water they take to run. Sonic’s like a personal water concierge that looks after your home – helping you to save money, protect your home, and reduce your environmental impact too. Coming soon to a sink near you!