AGL’s wind farm efficiency from Aveva software
If Dave Bartolo can pull off his latest project at AGL it will be, in his words, a miracle. And save the energy giant millions of dollars.
If Dave Bartolo can pull off his latest project at AGL it will be, in his words, a miracle.
Bartolo, who as AGL Energy’s head of asset intelligence has built a reputation around the world for being the vanguard of tech innovation in big energy companies, hopes to take an old hydro power station in northeast Victoria and make it a high-tech wonder.
To do it he’s using new tools and some old fashioned data entry to scan the hand-drawn plans for the Bogong plant and attach digital data to every item in it.
A digital thread of an entire power plant where anyone at AGL can access the reams of information on any item or component or manual ever associated with the site.
He hopes to have it done by June next year; the speed of such a task is made possible by combinations of new technology, led by software firm AVEVA, and Bartolo’s experience in digitalising energy assets.
Such is his reputation that at the AVEVA World showcase in San Francisco in October attended by more than 2000 people, Bartolo was sought out by others. “We are the canary in the coal mine. Every time I come to an event like this I get mobbed,” he says.
“(I get asked) what about this, are you having trouble with that? Australia is absolutely a test case for many companies.”
Getting to the position of trying to digitise an entire hydro electricity site comes after he got some wins on the board with wind turbines.
For wind farms to work really well, the veins of each turbine have to always be facing into the wind.
Sounds simple, but why and when it wasn’t happening was time consuming to find and fix.
In 2015, Bartolo started a pilot program for diagnostic and anomaly detection across the turbine fleet with 3500 models and 52,000 data points.
He now gets data every second, instead of in 10-minute intervals.
AGL discovered that as turbines age they become less efficient at doing the very thing they need to do.
It then built an algorithm that was able to spot, without needing to visit the turbine, whether that turbine was successfully facing into the wind.
In just three years AGL says it saved $18.7m in reduced forced outages and optimised maintenance. Across its business AGL estimates the software has saved it $50m to $70m in avoided failures.
AGL also estimates real-time data sharing with partners reduced time generating reports, saving an estimated 35,000 in annual AGL worker hours.
Bartolo is now more ambitious for his Bogong project.
“I know most of the drawings were hand written by Toshiba. So knowing that those drawings are challenged – just scanned bits of paper – if we can pull off the miracle to make them interactive and digital, using AVEVA software and some prep work, if I can demonstrate we can do it there, we can do it anywhere,” he says.
“What we believe is you’ll be able to look at a drawing of a turbine and find everything associated with that turbine close to instantly.
“That would be a pretty amazing experience.
AGL has been using AVEVA – an industrial software firm that offers technical solutions across energy, manufacturing and water, and smart city programs among others – since 2013.
The software firm, headquartered in the UK, has 200 staff across Australia, including software engineers working on some of its core products.
Its big push is to be an agnostic platform provider for its own and even competitor software programs.
It also showcased what it thinks will soon be a reality using large language AI chatbots to go a step further and allow, for example, an engineer in the control room who notices a drop in output to ask a chatbot what could be causing it.
AI then searches through reams of pages of manuals or documents for a possible fix. The worker can zoom in on the screen to a digital copy of the exact part on the turbine that needs attention and ask AI to send a summary of the issue to the maintenance department – all within minutes.
Bartolo says that idea was both exciting and confronting.
“We have prepared for chatbot. That was quite confronting,” he says.
“Once I saw that’s where we are headed, saw how AI will work in engineering … I realised
if I don’t have all my silos of data interactively connected, ready for a AI to look at, we’ll go nowhere.”
Tech to the rescue?
The rush to greener energy transition is also causing headaches Bartolo hopes technology can help solve.
“With energy transition – we are moving away from these big centralised assets where everything lives on the asset. We have assets that are seven hours drive away. We are going to get there once a year if we are lucky,” he says.
“Having a 3D experience, a global-mapped experience, a live interconnectivity experience it’s not going to be luxury soon, it’s going to be essential.
“And with AVEVA the sky’s the limit and that wasn’t there five years ago.
We have massive hurdles and barriers to overcome to do a full transition to renewables. A lot of people say it’s happening too slowly; actually it’s going incredibly fast because of the barriers we still need to invent solutions for.
“My view is quite balanced. I am optimistic – we definitely are going hard at this and we will do it, But it’s going to be a rough road and Damien Nicks has said that again and again. AGL has the largest generation portfolio in Australia.
“We have already shut down 32 per cent of our fossil fuel fleet in the last five years and it’s accelerating.
“We have to invent solutions in the meantime, so it’s both exciting and terrifying.”
Drones and AI
Outside the control rooms, AGL uses drones and technology to photograph and assess the entire steel cladding of Loy Yang and Bayswater for another pilot project.
“One of the problems we have is that the tech screws become corroded and can break off,” explains Bartolo.
If a sheet loses one or two screws, that’s fine but as more corrode the sheet becomes a potential hazard that can fly off and cause injuries.
The drone footage is being scanned by AI to map every single screw and every edge of cladding to provide a risk assessment of each piece.
“That’s exciting. Drones and AI coming together, because having 10 hours of footage of the cladding is useless in itself,” Bartolo says.
“It’s all that automatic assessment that makes it valuable and all the difference.
“We need to prepare all our data silos for an AI hit. And the cleaner the data, the better the outcome will be.
“I think the way history will judge is how well we prepare our data for AI to be able to do its magic.”
The writer was a guest of AVEVA at its showcase event in San Francisco in October.
More Coverage
