Fugro Roames has recruited drones to help scan 1.1 million overhead power poles across 160,000 kilometres for regional Queensland electricity supplier, Ergon Energy.
The former Ergon business unit – which was acquired by Dutch company Fugro in 2014 – has built an asset management service that has solved the challenges of maintaining the QLD energy supplier’s geographically-dispersed overhead electricity network.
Two Falcon unmanned aerial vehicles (drones) are being used in addition to Fugro Roames’ fleet of four Cessna 208 Caravan light aircraft. The drones are fitted with cameras and the Cessnas have laser sensors that capture 3D images of the distribution network that runs through 600 towns and cities across the state.
Drones, which fly much lower than the Cessna aircraft, enable the organisation to look a lot more closely at the top of the power poles to identify potential failures that may occur in 5 or even 10 years, Scott Carpenter, work program manager at Fugro Roames told CIO Australia.
Carpenter said he personally saw this type of failure during a trip to Brunswick Heads in northern NSW where a street light in a park was about to collapse because it had rusted through.
“That was exacerbated due to fat pelicans sitting on and landing heavily [on the pole] and the salt [in the air]. The arm was about to bend and that was a factor that was completely out of the control of [the energy provider] and they fixed it.
“I happened to be there on holidays, I rang up and they fixed it in the evening on a Friday night, totally unplanned with the crews working in the dark. While they are trained for that, we want remove people from potentially risky situations,” he said.
This could have been scheduled during the day and picking up this type of fault is what the drones are useful for, he said.
Before Roames was founded, Ergon spent over $100 million annually on inspecting, auditing and cutting vegetation that was obstructing overhead power lines. Since the introduction of the asset management service, costs have dropped to $60 million, said Carpenter.
“The secret sauce is all of that data goes into Amazon’s [data centre] and we use an algorithmic approach that is using machine learning to look into that 3D data and interpret it as a human does,” said Carpenter.
“We can build a model of the environment and run the analytics at scale, showing where poles are learning too much, where wires are hanging over the ground, and where has someone done a renovation that is too close to the power line.
“All of that [information] becomes very easy to harvest. Effectively, you are hiring an army of computers to inspect the entire network – you look at 10,000 locations and report 100 that need action," he said.
Sign up for Computerworld eNewsletters.