Utilities have always been heavily reliant on subject-matter expertise (SME) to drive asset maintenance work scheduling. Engineers leverage their industry knowledge of asset measurements and inspection results, but human beings are limited in what they can consider. A data-driven approach can enable SMEs to consider more signals from more assets to predict future asset health and recommend the best possible program for preventative maintenance.

Exelon's Infrastructure and Safety Analytics team explains how it developed a novel machine learning model to recommend preventative maintenance for transformers at the right time, to replace a time-based preventative maintenance program. They will explain the underlying on-premises and cloud analytics platform required to support this work and show how a data-driven culture, focused on cross-collaboration and diverse roles, enabled teams to work together to identify and solve problems.

Key Takeaways:

• Lessons learned in deploying data analytics platforms and managing a big data analytics platform (on-premise or cloud) with limited resources.

• A strong data-driven culture is important with a focus on cross-collaboration and support of diverse roles to work together to identify, ideate and solve problems.

• Investment is important during project ideation on adoption strategy.

• A novel machine learning model can recommend preventative maintenance for transformers at the right time, which can replace time-based preventative maintenance.