OG90
Course Contents
Module 1: Industrial Data Foundations and Analytics Strategy
Key Topics
- Industry context: big data, digital transformation and smart manufacturing trends
- Structure of plant data: SCADA records, DCS tags, historian databases and LIMS logs
- Core concepts: descriptive, diagnostic, predictive and prescriptive analytics levels
- Machine learning types: supervised, unsupervised and reinforcement learning paths
- Data engineering steps: handling missing records, sensor noise, signal spikes and outliers
- Data synchronization: sampling frequencies, time-stamps, averaging and lag variables
- Identifying high-value use cases, formulating project scopes and tracking business value
Module 2: Machine Learning Models and Performance Evaluation
Key Topics
- Linear and non-linear regression methods for continuous process outputs
- Classification models for predicting equipment states, operational modes and event flags
- Principal component analysis (PCA) and clustering tools for complex data compression
- Soft sensors: building virtual transmitters for difficult or delayed laboratory measurements
- Energy consumption, emissions profiles and utility system performance modeling
- Anomaly detection algorithms for identifying abnormal operating conditions early
- Model validation techniques, cross-validation loops, error metrics and interpretation boundaries
Module 3: Refinery and Gas Processing Applications
Key Topics
- Crude unit, FCC, hydrocracker and catalytic reformer process analytics
- Product blending optimization, quality giveaway reduction and laboratory property prediction
- Gas treating, glycol dehydration and acid gas removal column optimization models
- NGL recovery, demethaniser columns, fractionation trains and ethane recovery analytics
- Rotating and static asset tracking: compressor, pump, heat exchanger and furnace monitoring
- Network analysis workflows: managing flare gas, fuel gas headers, hydrogen loops and steam systems
- Practical case examples of real-world margin improvement, hydrocarbon recovery and reliability gains
Module 4: Predictive Maintenance and Advanced Control Integration
Key Topics
- Equipment failure prediction, degradation curves and remaining-useful-life (RUL) principles
- Vibration, temperature and electrical condition monitoring analytics for rotating equipment
- Static asset analytics: tracking exchanger fouling trends, catalyst deactivation and furnace coking
- Alarm analytics, fault tree mapping and early warning indicators for process upsets
- AI support for advanced process control (APC), model predictive control (MPC) and real-time optimization
- Dynamic constraint monitoring, operating envelope limits and safe boundary enforcement
- Human-in-the-loop decision-support dashboards and visualization practices for control rooms
Module 5: Deployment, Governance and Sustainable Value
Key Topics
- Technology architecture: comparing edge processing, on-premise servers and cloud computing options
- Model lifecycle workflows: version control, deployment tracking and management of change (MOC)
- Cybersecurity considerations for operational technology (OT) and information technology (IT) connectivity
- Model health tracking: drift detection, accuracy degradation checks and automated recalibration protocols
- Governance, accountability, data ownership and responsible AI practices in industrial sites
- Performance auditing: measuring benefits, user adoption and organizational change tracking
- Practical capstone workshop: formulating a scalable AI and machine learning roadmap for active plants