Real-Time Optimization of Chemical Process Models
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Real-time optimization of chemical processes in production facilities, such as steam methane reforming, is hindered by the complexity and time-consuming convergence of models, making it impractical to set control targets efficiently amidst variable demand and costs.
Innovation Solution
The use of semi-empirical process models based on mass, energy, and pressure balances, with adjustable parameters, allows for the determination of predicted production and consumption rates, enabling real-time optimization of process parameters to maximize profitability.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If complex process models are used for real-time optimization of chemical processes, then the accuracy of production targets is improved, but the computational time and complexity increase making it impractical for real-time control
Solution Approach 1:
The patent segments the complex process model into multiple simpler sub-models, each representing a specific unit operation or process section. This segmentation allows the overall optimization problem to be divided into smaller, computationally manageable parts that can be solved more quickly while maintaining accuracy.
Solution Approach 2:
The patent extracts and removes unnecessary complexity from the process models by identifying and eliminating redundant calculations and simplifying mathematical relationships. This extraction of essential elements maintains model accuracy while reducing computational burden.
2Reliability
If detailed process models with many parameters are used, then the reliability of optimization results is improved, but the model complexity and difficulty of implementation increase
Solution Approach 1:
The patent implements dynamic adjustment of model parameters based on operating conditions. The model complexity adapts to the specific situation, using simpler representations when appropriate and more detailed models when needed, thereby maintaining reliability while reducing overall complexity.
Solution Approach 2:
The patent changes the parameters used in the model based on operating conditions and data availability. By selecting appropriate parameter sets and adjusting model fidelity dynamically, the system maintains reliable optimization results while avoiding unnecessary complexity.
Data Source
AI summary
Method of controlling production of a plant incorporating one or more chemical processes in which products are produced through consumption of raw materials. In accordance with the method, current production rates of the products are computed by semi-empirical process models that are corrected through error corrections of actual production rates to produce corrected models. The production is then optimized using the corrected models to maximize the variable margin gained upon the sale of the products. The optimization yields targets that either directly or at least influence consumption of the raw materials. The raw materials are then introduced into the process or processes in accordance with the targets.