Posters

Abstract

The energy transition is on a critical path where innovation, speed, and adoption matter. Open-source technologies are a key contributor to collaboration on a global scale, technology transparency, and knowledge sharing and accumulation, all of which are critical to accelerating the expansion to clean and renewable energy sources. In an effort to improve safety and productivity and to reduce operational cost and carbon footprint for wind power generation industry, we have developed a solution using two open-source technologies: an AI/ML wind turbine health predictor sponsored by Databricks and the energy industry's standard OSDUTM Data Platform running on Amazon Web Services. Our collaborative effort used AI/ML to leverage consolidated diagnostics and sensor data to diagnose potential problem areas for wind turbines. By using the open-source OSDUTM Data Platform as the underlying data technology, we interchangeably shared the valuable findings among organizations, and across disciplines and business operations. As a result, not only was the solution able to detect and identify signals to predict wind turbine mechanical failure, it also intelligently delivered warnings and notifications to users. These notifications created awareness of and promoted actions toward reducing unplanned maintenance risk and costs for energy generation from wind turbines. With the prevention of wind turbine failure, the solution helped increase asset availability and simplify maintenance. This project is one example of companies coming together, accelerated by open-source technologies, and willing to contribute back to the larger community to drive further innovation and adoption. Wind power companies can use this system to study trends for turbine performance and build models for optimized proactive maintenance schedules. Our open-sourced solution with a reproducible predictive model means that anyone can leverage the resources to reduce cost, optimize maintenance, and improve safety for the wind power industry.