Hybrid Process Models Using Selective First-Principles Features
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current mathematical models used in industrial plants for predicting and controlling chemical reactions and processes lack interpretability and accuracy, particularly in real-time applications, due to their complexity and reliance on either first principles or data-driven approaches, which can lead to mistrust among operators and inefficiencies in plant operations.
Innovation Solution
The implementation of Reluctant Modeling principles, prioritizing measurable features over first principles augmented features, to create simpler and more interpretable Hybrid models that combine data-driven and first principle knowledge, ensuring models are based primarily on observable data with selective incorporation of domain knowledge.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If data-driven models are used to improve prediction accuracy, then model accuracy is improved, but interpretability deteriorates
Solution Approach 1:
The patent merges data-driven approaches with first principles knowledge to create hybrid models that maintain the predictive power of data-driven models while incorporating the interpretability of first principles. The system combines measured process data with fundamental chemical engineering principles to generate models that are both accurate and explainable to operators.
2Reliability
If first principles models are used to improve reliability, then model reliability is improved, but device complexity deteriorates
Solution Approach 1:
The patent extracts only the essential first principles knowledge needed for specific prediction tasks rather than implementing complete first principles models. The system selectively applies relevant fundamental principles (such as mass balance, energy balance, or specific chemical kinetics) to reduce model complexity while maintaining reliability for the target application.
3Measurement precision
If complex models are used to improve prediction accuracy, then prediction accuracy is improved, but ease of operation deteriorates
Solution Approach 1:
The patent segments complex models into modular components that can be independently understood and managed. The hybrid modeling approach divides the prediction task into parts handled by data-driven methods and parts handled by first principles, making the overall system more manageable and easier to operate while maintaining high accuracy.
Data Source
AI summary
Computer implemented methods and systems generate an improved predicted model of an industrial process or process engineering system. The model is a function of measurable features of the subject process and selected first principle features. First principle features are selected that capture linearities in a residual of a linear model constructed using a received dataset of the subject process. The model can further be a function of a scaled spline. The scaled spline is generated by computing a spine for a measurable feature of the subject process, fitting the computer spline to the residual of the constructed linear model, and scaling the fitting spline with a scaling factor. The model results in improved predictions of behavior of the subject process by relying primarily on the data of the measurable features of the subject process.