Plant Abnormality Prediction Using Correlation-Based Tag Selection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional plant abnormality prediction systems face inaccuracies due to improperly designed input and output tags, leading to imprecise predictions, especially when the understanding of the system is low, and fail to effectively analyze correlations between tags.
Innovation Solution
A plant abnormality prediction system that calculates correlation coefficients for tags, determines relevant and independent tags through clustering and filtering, and uses these determinations to perform accurate predictions using algorithms like k-NN, MLRM, or ensemble-based methods.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional tag selection methods are used without correlation analysis, then the model construction process is simple, but the prediction accuracy deteriorates due to inclusion of unrelated tags
Solution Approach 1:
The patent performs correlation analysis and clustering operations before model construction to pre-identify relevant and independent tags. This preliminary action filters out unrelated tags in advance, ensuring that only high-quality input tags are used in the subsequent prediction model, thereby improving prediction accuracy without requiring complex model structures.
Solution Approach 2:
The system automatically performs correlation analysis, clustering, and tag selection without requiring manual intervention or deep domain expertise. The algorithm self-identifies relevant tags and their relationships, making the complex tag selection process transparent and automated, thus improving accuracy while keeping the user-facing complexity low.
2Measurement precision
If tags with high correlation are included in the model, then the model uses more input information, but the prediction accuracy deteriorates due to redundancy and multicollinearity
Solution Approach 1:
The patent extracts and identifies tags with high correlation through clustering operations, then selectively removes redundant tags from the model. By taking out only the necessary independent tags and excluding highly correlated duplicates, the system maintains prediction accuracy while reducing the number of input tags to an optimal level.
3Loss of time
If the model is designed by someone with insufficient system understanding, then the development time is reduced, but the prediction accuracy deteriorates due to improper tag selection
Solution Approach 1:
The system performs automatic correlation analysis and tag selection without requiring the model designer to have deep system understanding. The algorithm self-identifies relevant tags and their relationships, making the complex tag selection process transparent and automated, thus improving accuracy while keeping the user-facing complexity low.
Solution Approach 2:
The patent replaces manual expert analysis and judgment with automated computational algorithms for correlation analysis and clustering. This substitution of mechanical expert reasoning with algorithmic processing enables non-experts to build accurate models quickly, reducing development time while maintaining or improving prediction accuracy through consistent, data-driven tag selection.
Data Source
AI summary
A system and method predict whether or not a plant is abnormal and perform an accurate prediction even if a modeling is executed in a state where the understanding for a target to abnormality determination is low, or when a person unfamiliar with system designs a prediction model. The system includes a correlation coefficient calculation unit for calculating a correlation coefficient for each of two tags among a plurality of tags; a relevant tag determination unit for determining a relevant tag for each tag of the plurality of tags by comparing the correlation coefficient with a reference value; and an independent tag determination unit for determining one or more among the plurality of tags as an independent tag based on the relevant tag. The relevant tag determination unit includes primary and second tag extraction sections for extracting primary and second tags for each tag of the plurality of tags.