Plant Data Classification with Adaptive Category Size Control
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods struggle to effectively visualize relationships among operation parameters, condition parameters, and evaluation indices in plant data, especially when the number of parameters exceeds three, and fail to model the relationship between plant states and evaluation indices accurately.
Innovation Solution
A plant data classification device that classifies multidimensional operation data into categories based on similarity using Adaptive Resonance Theory (ART), calculates evaluation indices, and adjusts category sizes to ensure variation is within a reference value, allowing for accurate modeling of relationships between plant states and evaluation indices through three-dimensional data visualization.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If data clustering technique is used to classify multidimensional operation data, then the number of measurement points can be handled, but it is difficult to visualize the relationships among parameters when the total number of items is three or more
Solution Approach 1:
The patent maps multidimensional operation data into a three-dimensional coordinate system where the X-axis represents operation parameters, the Y-axis represents condition parameters, and the Z-axis represents evaluation parameters. This dimensional mapping enables effective visualization of relationships among multiple parameters by transforming high-dimensional data into a visually interpretable 3D space, resolving the contradiction between handling multidimensional data and visualizing parameter relationships.
2Measurement precision
If ART categories are increased to improve classification accuracy, then more detailed categorization is achieved, but the complexity of the classification system increases
Solution Approach 1:
The patent dynamically adjusts the vigilance parameter of the ART network based on the variation of evaluation indices within each category. When the variation exceeds a threshold, the vigilance parameter is increased to create more refined categories, and when variation is acceptable, the parameter is decreased to reduce complexity. This adaptive parameter adjustment maintains classification accuracy while controlling system complexity.
Solution Approach 2:
The patent implements a feedback mechanism where the variation of evaluation indices within each category is continuously monitored. This feedback information is used to dynamically adjust the vigilance parameter and reclassify data when necessary, ensuring that the classification system adapts to maintain accuracy without unnecessary complexity.
3Measurement precision
If category size is decreased to reduce evaluation index variation, then classification precision is improved, but the number of categories increases
Solution Approach 1:
The patent makes the category size dynamic by adjusting the vigilance parameter based on the actual variation of evaluation indices. Instead of using fixed small categories, the system dynamically determines the appropriate category size needed to achieve acceptable evaluation index consistency, thereby improving precision without unnecessarily increasing the number of categories.
Data Source
AI summary
A plant data classification device according to one aspect of the invention includes: a data classification unit that classifies multidimensional operation data into categories; an evaluation index calculation unit that calculates an evaluation index of a category from a value of the operation data; a classification result evaluation unit that calculates a variation in the evaluation index for each category and determines whether the variation in the evaluation index is less than or equal to a reference value; and a parameter changing unit that changes, when it is determined that the variation in the evaluation index exceeds the reference value, a value of a parameter that defines a size of a category of the data classification unit in a direction of decreasing the size of the category.