Hybrid Manual Automated Data Mining Correlation Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing data analysis systems, both fully automated and manually directed, struggle to effectively determine the causes of abnormalities in multivariate models, particularly in large data sets, leading to inefficiencies in identifying equipment faults and necessitating a more effective monitoring system.
Innovation Solution
A process data analyzer is provided, comprising a computing system with a processor and memory, which allows users to select ranges of multivariate model output data and employs multivariate analysis techniques to determine the contributing process data, using methods like Partial Least Squares and AdaBoost to rank input contributions and identify potential faults.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If fully automated data analysis systems are used, then productivity is improved, but the ability to accurately determine causes of abnormalities deteriorates
Solution Approach 1:
The system segments the data analysis process into two distinct phases: an automated phase that quickly processes multivariate model outputs to identify potential abnormalities, and a manual expert phase that investigates and determines the root causes of identified abnormalities. This segmentation allows the system to leverage the speed of automation while preserving human expertise for accurate fault cause determination.
2Measurement precision
If strictly manual techniques are used, then measurement precision is improved, but productivity deteriorates
Solution Approach 1:
The system applies partial automation by using automated techniques only for the initial screening and identification of abnormalities in multivariate model outputs, rather than attempting to fully automate the entire analysis process. This partial application of automation maintains high productivity for the routine screening task while reserving manual expert analysis for the more complex and time-consuming fault cause determination, achieving optimal balance between speed and accuracy.
3Reliability
If multivariate models with dozens of inputs are used, then the ability to detect complex faults is improved, but the difficulty of determining causes of abnormalities increases
Solution Approach 1:
The system introduces an intermediary layer consisting of automated analysis tools that process the complex multivariate model outputs and present simplified abnormality indicators to operators. This intermediary layer translates the complex relationships among dozens of inputs into more manageable abnormality signals, reducing the cognitive burden on operators while maintaining the comprehensive fault detection capability of the multivariate models.
Data Source
AI summary
A method is provided for determining one or more causes for variability in data. The method includes selecting a first range of a multivariate model output data on a user interface and employing a computing system, operatively coupled to the user interface, to determine one or more process data causing a variability of the multivariate model output data in the first range when compared to a second range of the multivariate model output data. At least some of the process data includes data derived from a physical measurement of a process variable.