Queensland start-up using AI to detect faults in solar farms

Professor Rahul Sharma stands in front of solar farm PV panels (AI faults)
Professor Rahul Sharma

A tech start-up from The University of Queensland (UQ) is set to make efficiencies in the renewable energy sector by using artificial intelligence (AI) to detect faults in solar farm panels.

Associate Professor Rahul Sharma from UQ’s School of Electrical Engineering and Computer Science developed a system using machine learning algorithms to analyse data and detect faulty and underperforming solar panels, and to recommend targeted maintenance.

Related article: Pinpointing the exact location of power faults in minutes

Dr Sharma said the technology, SolarisAI, detects faults in solar farms without the need to install additional hardware, making it fast and cost effective.

“The challenge with large solar farms is detecting any faulty or underperforming solar panels hidden in a sea of millions,” he said.

“It’s impractical to install monitoring hardware on each panel, inspect every panel for damage or clean every panel to remove dirt.

“We needed to find a way to automate that process.”

Dr Sharma said the technology works at the array and string panel level and sequentially extracts vital information, monitoring for degradation, soiling, wiring faults and tracker problems, along with pinpointing any maintenance needed.

“Underperformance in Australian solar farms cost the industry around $400 million a year,” he said.

“We’re aiming for SolarisAI to reduce those losses by half, and potentially deliver an uplift in revenue of up to 8%.”

Discussions are underway to utilise the technology at Edify Energy’s Hamilton solar farms at Collinsville in North Queensland and Genex Power’s Kidston solar farm in North-West Queensland.

Edify Energy CEO and founder John Cole said it was an exciting project.

“The key to maintaining grid reliability and achieving success as a network operator is effective and efficient asset management,” he said.

Related article: Machine learning fast-forwards solar cell design

“This technology has the potential to drive solutions to the world’s energy crisis.”

The project partnered with German-based electronics and connection technology company Weidmuller to develop early prototypes.

Previous articleAVESS rebrands, reveals five-hour vanadium flow battery
Next articleGovernments announce joint funds for Pilbara Hydrogen Hub