Innovative Thinking  |  Mapping subsurface assets

Simple stories

An underground mapping and scanning technology which uses artificial intelligence is being piloted by National Grid in the United States, writes Belinda Smart.

Like doing a CT (computerised tomography) scan, MRI (magnetic resonance imaging) and ultrasound all at once.” It might sound like a new medical frontier, but this is a tagline for subsurface scanning technology Exodigo.

Exodigo is said to offer an important point of difference in subsurface scanning, which lies in “multi-sensor fusion”. In other words it combines several technologies in a “fusion” that optimises accuracy.

Exodigo combines several scanning and imaging technologies in a fusion and it is fitted onto a cart

These technologies include ground penetrating radar (GPR), electromagnetic surveying (EM), magnetic technology, metal detection and high-resolution 3D imaging as well as data from subsurface asset and object records.

All the data is stored in a cloud-based platform, supported by an artificial intelligence (AI) machine learning capability.

Exodigo strategic business development executive Bret Simon says Exodigo also has another differentiating feature from comparable technologies in the market. Rather than following the linear path typically used by subsurface scanning technology, Exodigo scans entire areas in what he describes as a unique “snakepath” criss-cross pattern using many transmitters to capture a high density of data points.

This results in the detailed and comparatively rapid scanning of terrain to uncover known and unknown subsurface assets. Data can be captured non-intrusively without disturbing surface objects, breaking ground or physically connecting to utilities by for example opening utility access holes.

The technology can be used by a single operator pushing a “cart”, no bigger than a lawnmower, to produce a clear picture of what is underground for stakeholders in energy, utilities, transport, construction and engineering.

It can also be fitted to drones.

The technology provides colour-coded schematics to identify where there are no subsurface objects, where definite objects have been identified and where there is an element of uncertainty.

Simon says another unique aspect of Exodigo is its machine-learning ability. Its algorithms are designed to enable it to teach itself, as he explains: “Say we produce a map and an engineer looks at that map which has been created for the first time and they can see an electric line, a gas line, a water line or a sewer line at particular GPS coordinates, created based on the sensor data from the GPR, EM, the metal detector, all the sensors.

“The next time Exodigo is switched on, it starts to learn and train and figure out that: ‘Okay, this is what an electric line looks like, this is what a sewer line looks like’.

“It’s training itself and that is a huge point of difference, the fact that it continues to expand its own capabilities.

“As we grow as a company this will only get faster and faster.”

He adds that he is optimistic that Exodigo will achieve widespread uptake because it can be quickly deployed to tell “simple stories” about complex subsurface areas.

“What we’re trying to get people to think is: ‘When you think about the underground, you should always just get an Exodigo map.’ It’s easy, it’s simple and it tells a story based on what you need.”

NATIONAL GRID

Exodigo has already grown rapidly. Founded in Israel in 2021, it now has offices in Tel Aviv and California. Last October, the company received a multi-million dollar investment – the exact sum was not disclosed – from National Grid Partners (NGP), the corporate venture and innovation arm of utilities giant National Grid.

In the second half of 2022, National Grid trialled Exodigo in the United States, with promising results that are being fed back to NGP.

National Grid engineering manager of gas innovation David Lessard explains that his team’s remit is the research and development, then potential deployment, of new technologies for National Grid’s gas business.

Lessard says Exodigo caught National Grid’s attention because capturing subsurface data still presents big challenges in the gas and utilities industries.

“The possibility of having a means to do that in a more efficient manner intrigued us,” he says.

Exodigo combines sensor technology with machine learning

“If we have records, we go out, we dig a hole. Typically, if it’s a congested area, we’ll use vacuum excavation or even hand digging, so that we’re not using a backhoe that could impact the assets. But it’s expensive to do that. Any time you excavate it’s expensive. So, if there’s a means to identify accurately where those assets are without digging, that’s a big efficiency.”

Another challenge is the uneven quality of existing subsurface asset records, he says. “Some of these assets are over one hundred years old and back then some of the mapping wasn’t as good, or those records have been lost. So that’s a major problem.”

Third-party contractors hitting underground assets also present problems. “If contractors are digging for a road construction or a fence, if anything hits our assets, that causes a major problem for us,” says Lessard.

PILOT PROJECTS

Lessard’s work evaluating Exodigo has involved recent pilots which took place at two sites in Long Island, New York State last year. One was at Garden City in Nassau at the western end of the island and the other at Yaphank in Suffolk County at the eastern end.

The Garden City site is a street scheme while Yaphank is a National Grid gas substation, a plot owned and operated by the organisation with known underground assets.

Scanning for the Garden City pilot ran from 17 to 29 August and the Yaphank pilot ran from 30 August to 1 September. Each pilot also included two weeks of analysis after the scanning work.

Lessard says the Yaphank site in particular provided an ideal blind test to see if Exodigo could find information “that we already know is there”.

“If there’s a means to identify accurately where those assets are without digging, that’s a big efficiency

He describes the results as “pretty successful” – a detailed underground map that validated the locations of all known assets, as well as two additional utility lines that had not appeared in existing records.

Evaluation of the Garden City site will be completed later this year, with additional testing being undertaken between April and November. Lessard says that site will furnish insights about using Exodigo in more built up, urban areas.

Overall, while National Grid is at an early stage of evaluating Exodigo, indications are promising, he says.

“The main application I see for it is for larger, complex projects, where we need to confirm where there are corridors that we can use in the design phase,” adds Lessard.

And he says he is particularly interested in evaluating how Exodigo works with the diversity of geography in the United States, particularly in congested urban environments.

Lessard concludes: “If we can get this to work in Brooklyn or Boston, it will be of immense value for us.”