Diagnostic Model Update Consistency via Master Data Comparison
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional state diagnostic devices for industrial machines face challenges in updating diagnostic models accurately and consistently, leading to inconsistencies in diagnostic results before and after updates, which affects the accuracy of abnormality detection.
Innovation Solution
A state diagnostic device that collects usage state and diagnostic data, selects master data from truly normal and abnormal states, and updates the diagnostic model only when the new model's results are consistent with the current model, ensuring no disruption in diagnostic accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If the diagnostic model is updated to improve diagnostic accuracy and increase diagnosable types of abnormality, then the diagnostic accuracy is improved, but the consistency of diagnostic results between before and after the update deteriorates
Solution Approach 1:
The system performs preliminary actions by collecting and storing master data representing truly normal and abnormal states before model updates. This pre-prepared reference data is used to evaluate new diagnostic models, ensuring that updates maintain consistency with established diagnostic standards while improving accuracy.
Solution Approach 2:
The system implements feedback by comparing diagnostic results from the new model against master data and existing diagnostic results. This feedback mechanism determines whether model updates are acceptable, preventing inconsistent updates while allowing accuracy improvements that maintain result consistency.
2Measurement precision
If machine learning is applied to update the diagnostic model, then the diagnostic accuracy is improved, but the complexity of the update process increases
Solution Approach 1:
The system extracts and separates the evaluation of model updates into distinct components: master data selection, diagnostic result comparison, and consistency determination. This extraction simplifies the complex update process by breaking it into manageable, independent steps that can be executed systematically.
Solution Approach 2:
The model update process is segmented into multiple discrete steps: collecting master data, training the new model, comparing diagnostic results, determining consistency, and finally updating or rejecting the model. This segmentation reduces overall complexity by making each step independent and manageable.
3Adaptability or versatility
If the diagnostic model is updated frequently to adapt to new abnormality types, then the adaptability is improved, but the risk of diagnostic inconsistency increases
Solution Approach 1:
Master data representing truly normal and abnormal states are collected and stored in advance, creating a reliable reference baseline before model updates. This preliminary preparation ensures that even as models adapt to new abnormality types, they are evaluated against established standards, maintaining reliability.
Solution Approach 2:
The system uses feedback from comparing new model results against master data and existing diagnostic results to determine whether updates are acceptable. This feedback loop ensures that adaptability improvements do not compromise the reliability of diagnostic results.
Data Source
AI summary
A state diagnostic device collects a usage state and diagnostic data of components, and selects, from the collected diagnostic data, at least one of diagnostic data obtained in a state where the usage state of the components is a truly normal state and diagnostic data obtained in a state where the usage state of the components is a truly abnormal state. The selected diagnostic data is defined as master data. A diagnosis result obtained by diagnosing the master data based on a current diagnostic model is compared with a diagnosis result obtained by diagnosing the master data based on a new diagnostic model, and whether the current diagnostic model is consistent with the new diagnostic model is determined. When the consistency is satisfied, the diagnostic model is updated from the current diagnostic model to the new diagnostic model.