Vibration Diagnosis Thresholding Using Current and Past Baselines
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing diagnosis systems face challenges in setting appropriate thresholds for diagnosis target information, such as vibration, especially when its tendency fluctuates under the same operating conditions, leading to difficulties in accurately diagnosing abnormalities.
Innovation Solution
A diagnosis apparatus that includes a sensor to detect vibration, a threshold setting unit that sets thresholds based on vibration data from a predetermined period, and a diagnosis unit that uses current and past thresholds to diagnose abnormalities, reducing the influence of external factors and noise.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If a threshold is set based on current vibration data only, then the diagnosis responds quickly to current state, but it is highly sensitive to external factors and noise causing false abnormalities
Solution Approach 1:
The patent applies preliminary action by calculating and storing past thresholds before they are needed for diagnosis. The system pre-processes vibration data to establish baseline thresholds from historical data, which are then used to compare against current vibration levels. This allows the diagnosis to filter out external factors and noise by referencing pre-established normal ranges, reducing false positives while maintaining quick response to actual abnormalities.
2Adaptability or versatility
If the threshold is set using a fixed time window, then the calculation is simple, but it cannot adapt to changing vibration tendencies over time
Solution Approach 1:
The patent implements dynamics by making the threshold calculation window adaptive rather than fixed. The system automatically adjusts the time window length based on the characteristics of vibration data and the specific diagnosis target. This allows the threshold calculation to adapt to changing vibration tendencies - using longer windows for stable conditions and shorter windows for rapidly changing conditions - while maintaining manageable computational complexity through automated window selection.
3Loss of information
If past thresholds are not considered in diagnosis, then the current state is evaluated in isolation, but trends and patterns over time are missed
Solution Approach 1:
The patent applies feedback by continuously comparing current vibration data against historically calculated thresholds and using the results to refine future threshold calculations. The system feeds back past threshold information to the current diagnosis process, creating a continuous learning loop. This allows the system to maintain accurate historical pattern recognition while managing complexity through systematic feedback mechanisms that automatically adjust to new data without requiring manual intervention.
Data Source
AI summary
A diagnosis apparatus includes: a sensor that detects diagnosis target information generated by a diagnosis target device; a threshold setting unit that sets a threshold for the diagnosis target information; and a diagnosis unit that diagnoses the diagnosis target device based on the diagnosis target information detected by the sensor and the threshold, in which the threshold setting unit sets the threshold based on the diagnosis target information in a predetermined period before a diagnosis time point, and the diagnosis unit performs diagnosis based on a current time point threshold set at the diagnosis time point and at least one past threshold set in the past.