Industrial Machine Diagnosis Using Abnormality Change Rate Alerts
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing diagnostic methods for industrial machines rely on threshold-based comparisons, which can be inadequate in environments where abnormality degrees drift due to environmental changes, leading to low sensitivity in detecting abnormalities and failing to distinguish between gradual and abrupt changes.
Innovation Solution
A diagnostic device that calculates an abnormality degree based on deviation from a reference state and includes a change rate calculation unit to assess the rate of change in the abnormality degree, allowing for alerts to be generated based on both the abnormality degree and its change rate.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Extent of automation
If a threshold is set for the abnormality degree to automatically issue warnings, then automation is improved, but sensitivity for detecting abnormalities deteriorates due to the need to provide margins for environmental changes
Solution Approach 1:
The patent applies dynamics by transitioning from a static threshold comparison to a dynamic evaluation that considers both the abnormality degree and its rate of change. The diagnostic device calculates the change rate of the abnormality degree over time and uses this dynamic information to determine whether to issue warnings, allowing the system to adapt to environmental changes while maintaining high sensitivity.
Solution Approach 2:
The patent changes the diagnostic parameters from simply comparing the abnormality degree against a fixed threshold to evaluating both the abnormality degree and its temporal derivative (change rate). This parameter expansion allows the system to distinguish between normal environmental drift and actual abnormalities, improving detection sensitivity while maintaining automation.
2Device complexity
If only simple comparison with threshold is used, then device complexity is reduced, but diagnostic accuracy deteriorates because it cannot distinguish between gradual and abrupt abnormality modes
Solution Approach 1:
The patent adds another dimension to the diagnostic evaluation by introducing the temporal dimension through change rate calculation. Instead of evaluating only the current abnormality degree (one dimension), the system evaluates both the abnormality degree and its rate of change (two dimensions), enabling differentiation between gradual and abrupt abnormality modes while maintaining reasonable system complexity.
3Reliability
If margin is provided to the threshold for environmental changes, then reliability is improved, but productivity deteriorates due to low sensitivity in detecting abnormalities
Solution Approach 1:
The patent implements feedback by continuously monitoring the change rate of the abnormality degree and using this information to adjust the diagnostic decision-making process. The system feeds back the temporal evolution information to the warning determination logic, enabling reliable detection under environmental changes while maintaining high productivity through automated dynamic evaluation.
Data Source
AI summary
The diagnosis device according to the present disclosure comprises: a data acquisition unit that acquires data indicating a predetermined state pertaining to industrial machine; a diagnosis unit that calculates an abnormality degree of the state on the basis of a separation degree of the data acquired by the data acquisition unit from a distribution of the data acquired in a reference state; a change degree calculation unit that calculates a degree of change in abnormality as a change degree; a first alert generation unit that compares the abnormality degree to an abnormality degree threshold value and determines whether or not a predetermined notification is required; a second alert generation unit that compares the change degree to a change degree threshold value and determines whether or not the predetermined notification is required; and a notification unit that outputs the predetermined notification on the basis of the results of the determinations by the first alert generation unit and the second alert generation unit.