The many facets of PdM: Oil and gas supermajor uses AI predictive analytics to improve efficiency and safety
The medley of predictive maintenance (PdM) strategies for improving machine health is growing larger and more powerful, whether using classic portable tools for non-critical asset inspection rounds and on-site problem verification and troubleshooting, or advanced technologies such as the IIoT, cloud, and AI and ML algorithms.
Leaders and analysts who go on record by documenting improvements gained from predictive maintenance initiatives provide a window into the immense potential of today’s enabling technologies. This article is one of seven diverse case studies that illustrate some of the many PdM methods and applications employed today.
The other case studies include:
- Analyst foresees AI/ML driving widespread adoption of prescriptive maintenance
- Midstream energy company uses IIoT strategy with integrated CMMS
- Consumer products manufacturer uses AI and ML models
- Self-driving truck company uses CMMS, BI tooling, and mobile app
- Tire manufacturer uses 24/7 wireless vibration monitoring system
- Thermal battery manufacturer uses Generative AI-driven data operations platform
- Mining company uses industrial edge data platform and SCADA system
Challenge: An oil and gas supermajor needed to maximize its production potential and improve the efficiency, availability, and overall safety of its offshore platforms. Thanks to previously completed digital transformation initiatives, its maintenance programs were already being operated with high digital competence and its assets were well instrumented and monitored. Its next target was adding predictive AI modeling capabilities.
Solution: An initial evaluation of Industrial AI Suite from SparkCognition, using a blind set of historical subsystem data from a key platform, exceeded expectations. AI models built for a separator system prone to unexpected failures predicted 75% of historical failures with an average of nine days advance warning. The models were then operationalized at scale at that platform and another. Industrial AI Suite was deployed in the supermajor’s remote onshore control center, providing alerting, 10-minute diagnostics, and a significant increase in overall operational visibility. AI predictive analytics were deployed across multiple critical subsystems to predict impending failures and optimize maintenance activities.
Results: The solution delivered a 4% increase in availability by avoiding net deferral events on both platforms. For instance, at least $10 million worth of deferred production was avoided when an alert enabled troubleshooting and scheduled maintenance of a critical export compressor with a faulty temperature sensor, which otherwise would have required up to two days to stage the asset to determine the root cause. The projected economic impact of Industrial AI Suite’s full deployment across the entire fleet of offshore platforms is roughly $800 million annually.
