Rotating Machine Abnormality Detection Using Frequency Spectrum Segmentation

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing rotating machine abnormality detection methods face challenges in accurately distinguishing between genuine abnormalities and noise, particularly when noise is non-reproducible and consistently present, leading to potential erroneous detection.

Innovation Solution

A rotating machine abnormality detection device that measures vibration caused by rotating bodies, processes frequency spectra to determine the presence of abnormalities, and uses noise determination based on frequency spectrum analysis to differentiate between noise-induced and genuine abnormalities, ensuring higher reliability in detection.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If noise filtering is applied to sensor output to remove noise, then measurement precision is improved, but when noise is non-reproducible and cannot be identified, erroneous detection occurs reducing reliability

Engineering Contradiction:

Improvenoise removal accuracyVSAvoidabnormality detection reliability

Core Design Contradiction:

Measurement precisionVSReliability

Solution Approach 1:

The patent segments the frequency spectrum into multiple frequency bands and independently analyzes the noise characteristics in each band. By dividing the overall frequency spectrum into discrete segments, the system can identify and filter noise in specific frequency ranges without affecting the detection of abnormalities in other frequency ranges, thereby improving both measurement precision and reliability

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The patent applies noise filtering selectively to specific frequency bands where noise is identified, rather than applying uniform filtering across the entire spectrum. This partial action approach removes noise from affected frequency ranges while preserving the integrity of abnormality signals in other frequency bands, preventing erroneous detection

Inventive Principle:

Principle #16Partial or excessive action

2Difficulty of detecting and measuring

If sensor output is analyzed to detect abnormality, then abnormality detection capability is improved, but noise mixed with sensor output causes erroneous detection reducing reliability

Engineering Contradiction:

Improveabnormality detection capabilityVSAvoiddetection accuracy

Core Design Contradiction:

Difficulty of detecting and measuringVSReliability

Solution Approach 1:

The patent applies different processing strategies to different frequency bands based on their local characteristics. Frequency bands identified as containing noise undergo noise filtering processing, while bands containing abnormality signals are preserved for detection. This localized quality approach ensures that each frequency band is processed according to its specific characteristics, improving both detection capability and reliability

Inventive Principle:

Principle #3Local quality

Solution Approach 2:

The patent performs preliminary identification of noise characteristics across frequency bands before conducting abnormality detection. By预先 identifying which frequency bands contain noise and which contain abnormality signals, the system can apply appropriate filtering in advance, preventing noise from causing erroneous detection while preserving genuine abnormality signals

Inventive Principle:

Principle #10Preliminary action

Applied Scientific Principles

This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.

Function Achieved in This Case

The solution enables more reliable abnormality detection by accurately distinguishing between noise and genuine abnormalities, reducing erroneous readings and improving the accuracy of rotating machine condition assessment.

Implementation Method 1

vibration caused by at least one of first and second rotating bodies is measured

Methodology Applied

Scientific EffectVibration: Vibration

Implementation Method 2

a predetermined feature amount related to a frequency component is obtained based on a frequency spectrum of the measurement data

Methodology Applied

Scientific EffectFrequency spectrum analysis:

Data Source

PatentUS11143632B2Rotating machine abnormality detection device and method, and rotating machine

Publication Date: 2021.10.12 KOBE STEEL LTD

AI summary

In a rotating machine abnormality detection device, a rotating machine abnormality detection method, and a rotating machine according to the present invention, vibration caused by at least one of first and second rotating bodies is measured, a predetermined feature amount related to a frequency component is obtained based on a frequency spectrum of the measurement data, and determination is made on presence/absence of abnormality based on the obtained predetermined feature amount. On this occasion, when determination is made that abnormality is present, determination whether the determination of abnormality is caused by noise or not is made based on the frequency spectrum, and when determination is made that the determination of abnormality is not caused by the noise, the determined abnormality is ultimately considered as abnormality.