Industrial Setpoint Optimization Using Historical Survival Matrices
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current process simulation modeling relies on theoretical equations that yield approximate values and are not effective for optimizing setpoints in industrial processes due to their inability to account for external variables and deviations from theoretical models.
Innovation Solution
A system utilizing a pseudo-process model based on historical data and survival matrices to optimize setpoints, which includes a historian server database, aggregation units, pseudo-process modeling units, and setpoint optimizer units, allowing for real-time and near-real-time adjustments without relying on first-principal equations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If theoretical equations are used to model process characteristics, then the modeling approach is simple and computationally efficient, but the accuracy of setpoint optimization is insufficient due to inability to account for external variables and operational deviations
Solution Approach 1:
The patent creates a pseudo-process model that copies the essential input-output relationships of the actual process without replicating the complex physical equations. Instead of modeling the underlying physics with first-principles equations, the system uses historical operational data to create an empirical model that replicates the process behavior, thereby achieving high accuracy without the computational burden of complex theoretical models
Solution Approach 2:
The system transforms the modeling approach by changing from fixed theoretical parameters in first-principles equations to dynamic empirical parameters derived from historical data. The pseudo-process model uses statistical parameters and relationships learned from actual process operation, allowing it to adapt to real-world variations and external factors that theoretical models cannot capture
2Reliability
If first-principles equations are used for setpoint optimization, then the computational resources are saved, but the reliability of the optimization is reduced due to deviations from theoretical models under operational conditions
Solution Approach 1:
The system performs preliminary analysis by pre-processing historical operational data to establish the pseudo-process model before actual optimization is needed. The model training and validation are done in advance using accumulated historical data, so that during real-time operation, the system can quickly query pre-computed relationships rather than performing complex calculations, thus achieving both high reliability and computational efficiency
Solution Approach 2:
The pseudo-process model serves itself by using its own stored empirical relationships to determine optimal setpoints without requiring external theoretical models. The system leverages its previously learned data-driven relationships to autonomously optimize setpoints, eliminating the need for computationally intensive first-principles calculations while maintaining high reliability
3Productivity
If theoretical models are used to determine optimal setpoints, then the system is easier to implement, but the productivity of the industrial process is not maximized due to inability to account for external environmental factors
Solution Approach 1:
The system implements feedback by continuously using historical operational data that includes actual process outcomes and performance metrics. The pseudo-process model is trained on this feedback-rich data, allowing it to learn from past successes and failures, and thereby optimize setpoints that actually improve productivity rather than just following theoretical predictions that may not reflect real-world constraints and opportunities
Data Source
AI summary
This disclosure is directed to a system for determining optimum setpoints for equipment in an industrial process. In some embodiments, the system does not use first-principles models to determine ideal setpoints. Instead, the system uses actual historical data and determines the setpoints at which the highest and/or longest key performance indexes were achieved according to some embodiments. In some embodiments, the system is able to save computer resources by reducing processing power through the use of a survival matrix as opposed to an iterative model. In some embodiments, the survival matrix is derived from statistical calculations on the historical data for KPI achieved timeframes.