Chemical Process Prediction Using Reaction Rate Modeling

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Improving prediction accuracy in processes involving reactions in chemical plants is challenging due to nonlinearity influenced by factors like temperature.

Innovation Solution

Generating a prediction model using reaction rate values, such as integral values calculated via the Arrhenius equation, and employing a hierarchical structure with multiple prediction equations to correlate process data, and using sampling methods to align data acquisition with response variables.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If conventional prediction methods (peak value, difference value, integral value) are used for chemical plant processes, then the prediction system can be implemented, but prediction accuracy deteriorates due to nonlinearity of reactions

Engineering Contradiction:
Improveprediction accuracyVSAvoidprediction reliability under nonlinear conditions
Core Design Contradiction:
Measurement precisionVSReliability

Solution Approach 1:

The patent transforms the prediction approach by changing from using conventional parameters (peak value, difference value, integral value) to using reaction rate values. This parameter transformation allows the prediction model to capture the nonlinear characteristics of chemical reactions more accurately, thereby improving both prediction accuracy and reliability simultaneously

Inventive Principle:
Principle #35Parameter changes

Solution Approach 2:

The patent performs preliminary calculation of reaction rate values using the Arrhenius equation before feeding data into the prediction model. This preliminary action of pre-processing the data with scientifically-grounded calculations enables the model to work with more meaningful inputs, improving prediction reliability without sacrificing accuracy

Inventive Principle:
Principle #10Preliminary action

2Measurement precision

If reaction rate values calculated via Arrhenius equation are used, then prediction accuracy improves, but device complexity increases due to additional calculations

Engineering Contradiction:
Improveprediction accuracyVSAvoidcomplexity of prediction system
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The patent introduces reaction rate values as an intermediary parameter between raw process data and the final prediction model. This intermediary layer, calculated through the Arrhenius equation, serves as a bridge that transforms complex nonlinear reaction behavior into a form that the prediction model can process effectively, improving accuracy while keeping the added complexity manageable

Inventive Principle:
Principle #24Intermediary (Mediator)

Solution Approach 2:

The patent replaces complex mechanical or empirical prediction mechanisms with a scientifically-based calculation approach using the Arrhenius equation. This substitution uses fundamental chemical principles to calculate reaction rates, providing a more accurate foundation for predictions without requiring overly complex system architecture

Inventive Principle:
Principle #28Mechanics substitution (Replace mechanical system)

Applied Scientific Principles

This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.

Function Achieved in This Case

Enhances the accuracy of predicting product characteristics in chemical plants by correlating process data effectively, allowing for better control and optimization of operating conditions.

Implementation Method 1

The value corresponding to the reaction rate may be calculated by using an Arrhenius equation

Methodology Applied
Scientific EffectArrhenius equation:

Data Source

PatentEP4102322B1Prediction device, prediction method, and program
Publication Date: 2025.08.27 DAICEL CORP
  • EP4102322B1 patent drawingFigure 1
  • EP4102322B1 patent drawingFigure 2
  • EP4102322B1 patent drawingFigure 3

AI summary

To improve prediction accuracy of a process including a reaction in a chemical plant. A prediction apparatus includes a process data processing unit that performs a predetermined processing process on process data obtained from a chemical plant, and a prediction model generation unit that generates a prediction model having learned features of the process data obtained from the chemical plant, on the basis of causality information that defines a combination of first process data and second process data or a value corresponding to the second process data among the process data obtained from the chemical plant or the process data processed by the process data processing unit. The first process data is used as an explanatory variable. The second process data or the value corresponding to the second process data is used as a response variable. Furthermore, the process data processing unit obtains a value corresponding to a reaction rate of a processing target in a predetermined period by using the process data.