Electric Motor Structural Fault Detection Using Temporal Symptom Vectors

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing methods for detecting anomalies in electric motor operation, such as motor current signature analysis, are less effective during transient conditions, making it difficult to differentiate fault signals from variations due to changes in operating conditions.

Innovation Solution

A method involving the acquisition of a measurement signal representing motor rotation, application of band-pass filters to extract frequency components, computation of a frequency signature vector, and generation of a temporal symptom vector, which is then used with a classifier model to detect structural faults, allowing for improved robustness and adaptability to changing conditions.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If motor current signature analysis is used for fault detection, then detection capability under steady state conditions is improved, but detection accuracy deteriorates during transient conditions

Engineering Contradiction:
Improvefault detection accuracyVSAvoidadaptability to transient conditions
Core Design Contradiction:
Measurement precisionVSAdaptability or versatility

Solution Approach 1:

The patent applies dynamics by making the filter bank adaptive to changing operating conditions. The central frequencies of the band-pass filters are dynamically adjusted based on the instantaneous motor speed, allowing the system to track frequency variations during transient conditions while maintaining effective fault signal extraction

Inventive Principle:
Principle #15Dynamics

Solution Approach 2:

The patent changes the parameters of the filtering system by using a bank of band-pass filters with varying central frequencies and bandwidths. These parameters are adjusted according to the operating condition (steady state or transient), enabling the system to optimize its detection capability for different operational phases

Inventive Principle:
Principle #35Parameter changes

2Reliability

If frequency domain analysis is used for fault detection, then fault signals can be identified under steady state conditions, but the method becomes less effective during transient phases with changing operating conditions

Engineering Contradiction:
Improvedetection reliabilityVSAvoidrobustness to operating condition changes
Core Design Contradiction:
ReliabilityVSAdaptability or versatility

Solution Approach 1:

The system dynamically adapts to changing operating conditions by adjusting the filter characteristics based on instantaneous speed. This dynamic adaptation maintains detection reliability during transient phases where traditional fixed-frequency methods fail

Inventive Principle:
Principle #15Dynamics

Solution Approach 2:

The frequency spectrum is segmented into multiple band-pass filters with different central frequencies. This segmentation allows the system to isolate and analyze specific frequency components associated with faults while filtering out unrelated variations, improving reliability across different operating conditions

Inventive Principle:
Principle #1Segmentation

Data Source

PatentEP4277113B1Method for detecting a structural fault of an electric motor
Publication Date: 2025.03.26 SCHNEIDER TOSHIBA INVERTER EUROPE SAS
  • EP4277113B1 patent drawingFigure 1~2
  • EP4277113B1 patent drawingFigure 3~4
  • EP4277113B1 patent drawingFigure 5~6

AI summary

The invention relates to a method for detecting a structural fault of an electric motor (1), comprising the steps: (i) acquiring a measurement signal (S) of a physical parameter representative of the rotation of the electric motor (1), (ii) obtaining the high frequency component of the measurement signal (S), resulting in a corrected signal (S_cor), (iii) Applying a set of band-pass filters (F-1, ..., F-n) to the corrected signal (S_cor) resulting in a set of filtered corrected signals (FS_cor-1, ..., FS_cor-n), (iv) Determining a stator frequency (Fs) and a rotor frequency (Fr), (v) Computing a frequency signature vector (FSV), (vi) Computing a temporal symptom vector (TSV) from the frequency signature vector (FSV) and the set of filtered corrected signals (FS_cor-1, ..., FS_cor-n), (vii) Detecting a structural fault of the electric motor (1) from the determined temporal symptom vector (TSV) and from a classifier model.