Rajib Saha

Each maintenance approach, predictive, condition‑based, and risk‑based, offers distinct strengths but also clear limitations. Predictive tools can surface early warnings but often create false positives; condition‑based monitoring improves visibility but requires the right sensing and data quality; and risk‑based inspection reduces turnaround scope but depends on consistent methodology and asset knowledge.

This workshop outlines a practical method for combining these approaches into a balanced, value‑driven maintenance mix that improves reliability, controls cost and enables standardization across sites.

Topics to be covered include:

• Defining maintenance objectives and success metrics: Starting with the why — what do you want to achieve?
• Determining AI KPIs: cost efficiency, reliability uplift, turnaround reduction, asset‑health indicators
• Maturing risk‑based inspection: reducing turnaround scope and optimizing inspection timing
• Selecting the most high‑impact use cases: ranking and then planning your roadmap
• Evaluating predictive approaches: physics‑based vs AI ML based
• Bridging the gap between detection and action: workforce adoption and reducing false positives
• Digitalization and standardization for scalability across refineries and operating sites