Diagnostic Analytics for Operational Excellence

about
A Multi-Site Renewable Energy Company
Problem
  • Impossible to review the millions of records gathered by their SCADA supervisory system.
  • Among these records are valuable insights about operational failures that are difficult to identify with so much data.
  • Hundreds of highly complex signals hide the genuine root cause of failure, making it difficult for humans to identify.
Action
  • Built Machine Learning (ML) and advanced statistics models for data driven diagnostic analysis
  • Developed powerful visualizations to tell the story of what happened and to reveal the root causes of failures with efficiency and clarity
  • Trained management for continuous re-training to increase efficiency and accuracy over time
Tool Stack
  • Python, Azure (Data Factory, Machine Learning Studio, Power BI)

100

+

Signals Processed

35

B+

Billions of Records

90

%+

Accuracy of Root Cause Detection