Dynamic Blending Element Models for Accurate Product Prediction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current systems face challenges in accurately predicting the characteristics of a blended product in processing plants, as they often rely on volume-based determinations which are inadequate for modeling future properties of blended inputs.
Innovation Solution
A computer-implemented method using dynamic element models such as pool tank, non-buffer, and heel volume models to generate predicted blending elements, which can be applied to plant-wide optimization and control of physical components, enabling the production of optimized product distributions based on target time intervals.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If volume-based determinations are used to predict blended product characteristics, then the modeling process is simple, but the prediction accuracy is insufficient
Solution Approach 1:
The patent applies dynamic element models that account for time-varying characteristics of blending processes. Instead of static volume-based calculations, the system uses dynamic models (pool tank, non-buffer, heel volume) that capture the temporal evolution of blended product properties, thereby improving prediction accuracy while maintaining computational tractability through structured model formulations.
Solution Approach 2:
The patent transforms the modeling approach by changing from simple volume parameters to comprehensive dynamic parameters including flow rates, blending element concentrations, and time-dependent properties. This parameter transformation enables accurate prediction of future blended product characteristics by incorporating multiple influencing factors beyond just volume measurements.
2Measurement precision
If dynamic element models are used to accurately predict blending characteristics, then prediction accuracy improves, but computational complexity increases
Solution Approach 1:
The patent segments the blending process into distinct dynamic models based on tank configuration and operating conditions (pool tank model for well-mixed tanks, non-buffer model for inline blending, heel volume model for tanks with residual material). This segmentation allows accurate prediction for each specific scenario while keeping individual model computations manageable by addressing only the relevant dynamics for each case.
Solution Approach 2:
The patent introduces blending elements as intermediary variables that mediate between input stream compositions and output product characteristics. These blending elements serve as state variables that capture the essential dynamics of the blending process, enabling accurate prediction of future product properties without requiring complex direct calculations from all input parameters at all times.
3Measurement precision
If future values of blending elements are modeled accurately, then product characteristic prediction improves, but data processing requirements increase
Solution Approach 1:
The patent performs preliminary calculation of blending element trajectories using the dynamic models to predict future product characteristics before actual blending occurs. By computing predicted blending element values in advance based on current state and input stream specifications, the system enables forward-looking product quality assessment and optimization without requiring extensive real-time data processing during the blending operation itself.
Data Source
AI summary
Embodiments of the present disclosure provide for improved modeling of blending elements. Such embodiments utilize particular dynamic element model(s) that accurately generate value(s) for predicted blending element(s) that function in accordance with any of a myriad of properties. Such dynamic element model(s) accurately account for changes in the blending element(s) over time as caused by any of a myriad of factors, including flow rate(s), tank volume(s), and/or heel volume(s). Some example embodiments generate at least one predicted blending element associated with a tank utilizing a pool tank dynamic element model, a non-buffer dynamic element model, or a heel volume dynamic element model, and output the at least one predicted blending element.