Broken-Rotor Bar Detection in Inverter-Fed Induction Motors

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing methods for broken-bar fault detection in inverter-fed induction motors are ineffective due to varying speed and load conditions, making it difficult to extract fault signatures using conventional motor current signature analysis (MCSA).

Innovation Solution

A system and method utilizing graph-based signal processing techniques to analyze stator current under varying speed and load conditions, employing sparsity and smoothness constraints to detect broken-bar faults by transforming stator current into a complex space vector, performing Short-Time Fourier Transform (STFT) and graph-based optimization to identify sparse and smooth frequency components in the spectrogram matrix.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If conventional MCSA methods are used for broken-bar fault detection, then detection is effective under stationary conditions, but detection performance deteriorates under varying speed and load conditions

Engineering Contradiction:
Improvefault detection accuracyVSAvoidapplicability under varying speed and load
Core Design Contradiction:
Measurement precisionVSAdaptability or versatility

Solution Approach 1:

The patent applies dynamics by transforming the stator current signal into a complex space vector and then into a time-frequency domain representation using STFT. This dynamic transformation allows the fault detection system to adapt to varying speed and load conditions by capturing the time-varying characteristics of the current signal, rather than relying on stationary frequency analysis.

Inventive Principle:
Principle #15Dynamics

Solution Approach 2:

The patent changes the analysis parameters by imposing sparsity and smoothness constraints on the fault signature extraction process. These parameter changes enable the system to distinguish fault components from normal operational variations under varying speed and load conditions, improving detection accuracy in dynamic environments.

Inventive Principle:
Principle #35Parameter changes

2Ease of manufacture

If short measurement periods are used for MCSA, then implementation is simple, but detection reliability decreases under variable operating conditions

Engineering Contradiction:
Improveimplementation simplicityVSAvoiddetection reliability under variable conditions
Core Design Contradiction:
Ease of manufactureVSReliability

Solution Approach 1:

The patent applies preliminary action by pre-defining sparsity and smoothness constraints before the actual fault detection process. These pre-established constraints guide the fault signature extraction, enabling reliable detection under variable conditions without requiring complex real-time adjustments or extended measurement periods.

Inventive Principle:
Principle #10Preliminary action

3Use of energy by moving object

If inverter-fed operation is used for efficiency, then energy efficiency improves, but fault detection becomes difficult due to varying speed and load

Engineering Contradiction:
Improveenergy efficiencyVSAvoidfault signature extraction difficulty
Core Design Contradiction:
Use of energy by moving objectVSDifficulty of detecting and measuring

Solution Approach 1:

The patent substitutes traditional mechanical frequency analysis with a time-frequency domain approach using complex space vector transformation and STFT. This substitution enables effective fault detection in inverter-fed motors by capturing the non-stationary characteristics of the current signal that arise from varying speed and load operations.

Inventive Principle:
Principle #28Mechanics substitution (Replace mechanical system)

Data Source

PatentEP4402495B1System for broken-rotor bar detection in inverter-fed induction motors using graph-based motor current signature analysis
Publication Date: 2025.06.25 MITSUBISHI ELECTRIC CORP
  • EP4402495B1 patent drawingFigure 1
  • EP4402495B1 patent drawingFigure 2
  • EP4402495B1 patent drawingFigure 3

AI summary

A computer-implemented method is provided for detecting broken bar faults of an induction motor during operations. The method includes steps of acquiring, in a time domain, a signal of a stator current powering the induction motor with a fundamental frequency via an interface, wherein the induction motor is under a varying speed operation; converting the stator current to a complex space vector; transforming the complex vector to a transformed stator current by referencing to synchronous reference frame; performing Short time Fourier Transform (STFT) on the transformed stator current to get spectrogram matrix; removing a DC component from the spectrogram matrix; determining, in a frequency domain, sparse and smooth frequency components in the spectrogram matrix, wherein the determining includes a graph-based method by imposing a smoothness constraint and a sparsity constraint on the frequency component; and detecting a fault in the induction motor if the frequency component includes a continuously changing and sparse fault frequency component in the spectrogram matrix.