A major refinery operator with 3 million BPD capacity across 36 plants, including a Delayed Coker unit.
Challenge
Geysers/steam eruptions during the “Coke Cutting” step in Delayed Coker Unit operations, spreading steam, hot water, and coke particles across a 2-3 mile radius.
Objective
Predict and minimize dangerous steam eruptions to enhance safety during coke cutting.
Our Solution Approach
Process Point implemented a data-driven intelligence solution to enhance operational efficiency and safety. Our approach involves meticulous data collection, advanced model development, seamless implementation, and continuous improvement to ensure optimal performance.
Data Collection and Analysis: Gathered historical data on eruption factors and applied Principal Component Analysis to identify key variables.
Model Development: Built a predictive model using machine learning, achieving 93% prediction accuracy.
Implementation: Integrated real-time monitoring and developed operational guidelines.
Continuous Improvement: Established ongoing monitoring and model refinement.
Process Point’s data-driven approach successfully mitigated a critical safety risk while boosting operational efficiency. This project showcases our ability to leverage advanced analytics in solving complex industrial challenges, resulting in safer and more profitable operations for our clients. The implementation of continuous improvement strategies ensures the solution’s long-term effectiveness and adaptability to changing conditions.
