Vibration Analysis Using Dual Mahalanobis Spaces for Failure Timing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing vibration analysis systems fail to predict the abnormality generation period of machines, limiting effective predictive maintenance.
Innovation Solution
A vibration analysis system that calculates Mahalanobis distances in multiple signal spaces to predict abnormality generation periods by analyzing vibration signals, using a first unit space of signal intensities and a second unit space of gravity center data, with threshold-based warnings for different timeframes.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If vibration signals are analyzed using conventional proximity methods to determine abnormality presence, then abnormality detection capability is improved, but the ability to predict abnormality generation period is lost
Solution Approach 1:
The patent segments the vibration signal analysis into multiple independent Mahalanobis distance calculations, each operating on different signal spaces (first signal space with signal intensities, second signal space with gravity center data). This segmentation allows simultaneous abnormality detection and prediction capability, as each space provides different predictive insights without interfering with the other.
Solution Approach 2:
The patent introduces a new dimension of analysis by calculating Mahalanobis distances in multiple signal spaces simultaneously. The first Mahalanobis distance operates on signal intensities across frequency bands, while the second operates on gravity center data derived from the same signals. This multi-dimensional approach enables both detection and prediction functions to coexist and complement each other.
2Reliability
If multiple Mahalanobis distance calculations are performed to enable abnormality prediction, then predictive maintenance capability is improved, but system complexity increases
Solution Approach 1:
The patent makes the Mahalanobis distance calculation mechanism universal by applying it to multiple signal spaces with different purposes. The same mathematical framework serves both abnormality detection (first signal space) and prediction (second signal space with gravity center data), eliminating the need for separate complex systems for each function.
Solution Approach 2:
The patent implements feedback through the gravity center calculation portion, which continuously tracks the center of gravity of signal intensities across frequency bands. This feedback mechanism provides evolving predictive information about abnormality development over time, enabling proactive maintenance decisions while using a unified computational approach.
3Measurement precision
If detailed vibration signal analysis is performed to predict abnormality periods, then maintenance timing accuracy is improved, but processing time increases
Solution Approach 1:
The patent extracts key predictive features from the complex vibration signals by calculating signal intensities in specific frequency bands and deriving gravity center data from these intensities. This extraction process identifies the most informative aspects of the vibration data for prediction purposes, reducing processing requirements while maintaining high timing accuracy.
Solution Approach 2:
The patent transforms the vibration signal analysis by changing parameters - converting raw vibration signals into signal intensities across frequency bands, then into gravity center positions. These parameter transformations create simplified representations that are easier and faster to process while preserving the essential predictive information about abnormality development.
Data Source
AI summary
A vibration analysis system includes: a signal input portion that receives an input of a vibration signal detected by a sensor; an intensity calculation portion that calculates a plurality of signal intensities corresponding to a plurality of frequency bands by analyzing the vibration signal; a first distance calculation portion that calculates a first Mahalanobis distance of a first signal space configured of the plurality of signal intensities with respect to a first unit space; a gravity center calculation portion that calculates two-dimensional gravity center data indicating gravity center positions of the plurality of signal intensities; a second distance calculation portion that calculates a second Mahalanobis distance of a second signal space configured of the gravity center data with respect to a second unit space; and an abnormality prediction portion that predicts an abnormality generation period of the object based on the first Mahalanobis distance and the second Mahalanobis distance.