Hybrid Process Modelling: Driving Improved Performance
Integrated AI and Data Intelligence for Enhanced Plant Operations.
Why Hybrid Process Modelling?
First-Principles Models
Based on engineering principles, process simulations, and thermodynamics models
- However, doesn’t capture all phenomenon
- Susceptible to errors and doesn’t reflect operating plant
AI/ ML
Models
Based on core data science principles and AI-ML modelling
- Big Data requirements
- Extrapolation issues on violating physical constraints
- Difficult to interpret if inferior data quality
Hybrid Process Modelling
First-principles foundations complement data-based AI predictions for precision!
- Accurately simulates complex processes in plants
- Improves predictive and prescriptive insights
- Self-sustaining Fit-for-Purpose models
Different Types of Hybrids & What Suits
Your Plant/Process
- AI Dominant Hybrid
- Simulated AI Hybrid Models
-
First Principles Dominant
Hybrid Models
- Applied to complex reaction unit operations, new material processes, and new technology processes such as the Hydrocracker reactor model
- AI/ML models are developed based on plant/experimental data, augmented with first principles, process constraints, and domain knowledge
- Generates entirely new, more accurate predictive models, enabling full-scale AI deployment across plant operations
- Applied for value chain-wide models from crude oil input to finished chemical output, building rapidly deployable and compact models online at the edge.
- AI/ML model built using simulation data, constraints and domain expertise to create a high-fidelity, purpose-fit solution that is accurate within its trained range and fully democratized for use.
- Users can easily extend the scale of modeling from units to the entire site and synchronize the model across design, operations and maintenance.
- Augments an existing first principles model with AI, using data from operations to calculate unknown variables and relationships not captured by the original model.
- AI / ML determines the unknown value and its relationships to continuously calibrate the model as conditions change.
- This is a natural extension to existing first principles models in brownfield deployments; it is quick and easy to adopt and significantly increases accuracy.
AI Dominant Hybrid
Simulated AI Hybrid Models
First Principles Dominant Hybrid Models
AI Dominant Hybrid
- Applied to complex reaction unit operations, new material processes, and new technology processes such as the Hydrocracker reactor model
- AI/ML models are developed based on plant/experimental data, augmented with first principles, process constraints, and domain knowledge
- Generates entirely new, more accurate predictive models, enabling full-scale AI deployment across plant operations
Simulated AI Hybrid Models
- Applied for value chain-wide models from crude oil input to finished chemical output, building rapidly deployable and compact models online at the edge.
- AI/ML model built using simulation data, constraints and domain expertise to create a high-fidelity, purpose-fit solution that is accurate within its trained range and fully democratized for use.
- Users can easily extend the scale of modeling from units to the entire site and synchronize the model across design, operations and maintenance.
First Principles Dominant Hybrid Models
- Augments an existing first principles model with AI, using data from operations to calculate unknown variables and relationships not captured by the original model.
- AI / ML determines the unknown value and its relationships to continuously calibrate the model as conditions change.
- This is a natural extension to existing first principles models in brownfield deployments; it is quick and easy to adopt and significantly increases accuracy.
