Harnessing big data, analytics and AI for a green energy game-changer
Enbridge’s tech-forward solution could ‘fundamentally transform’ economics, efficiency of wind power
Big data. Machine learning. Predictive analytics.
Enbridge, a Canadian leader in green energy investment, is using all of these elements to help optimize our wind power performance across North America.
Enbridge’s one-of-a-kind Performance Analytics and Situational Awareness (PASA) solution could be a game changer in a sector grappling with increasing maintenance costs. PASA helps optimize turbine servicing, avoid downtime, and predict mechanical issues—and could one day help ensure the viability of the wind energy industry as a whole.
“We believe the application of machine-learning techniques, as evidenced in our PASA solution, will fundamentally transform the economics and efficiency of wind power,” says Tony Khoo, Enbridge’s Manager of Advanced Analytics, whose team developed and delivered the PASA solution.
“When our team first tackled this issue in late 2015, we knew that maintenance was the largest cost for wind power operations—and, potentially, a barrier to investment. However, there were no predictive maintenance programs out there at the time with the reliability, sensitivity or speed to create meaningful industry change,” says Khoo.
“The PASA solution has helped Enbridge evolve our wind energy maintenance program to one that’s more productive, trims costs, and advances renewable energy production for a greener future.”
To date, Enbridge has committed more than $7.8 billion in capital to renewable energy and power transmission—including 18 onshore and offshore wind farms in North America and Europe with more than 3,400 megawatts (MW) of gross capacity, based on projects either in operation or under construction.
To create the PASA solution, which has been fully operational since early 2017, Enbridge’s advanced analytics team developed machine-learning models to identify wind turbine blade defects, and estimate time-to-failure and remaining useful life for any given turbine blade.
Data is currently being streamed from 2,460 turbine blades, on 820 turbines, at 11 Enbridge wind farms, and displayed as real-time operational dashboards for maintenance staff. To date, this solution has:
- Allowed us to predict, identify and act on wind turbine defects in a timely fashion, with 50 percent projected annual savings on inspection costs;
- Reduced blade inspection process time from three months to three weeks, a change that trims costs, boosts productivity and enhances safety; and
- Reduced maintenance budget errors by more than 20 percent.
“This innovation could be revolutionary for the wind energy industry, and one day set the benchmark for productivity of wind power programs,” says Khoo.
Enbridge researchers have been sharing their findings with industry and academic peers to help the industry improve the commercial viability of green energy projects.
“We’ve created new efficiencies through technological innovation,” says Khoo, “In doing that, we’re directly addressing the economic challenges of renewable energy—and, at the same time, furthering Canada’s green energy and reduced emission goals.”