Load Abnormality Detection Device for Induction Motors

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Conventional facility abnormality diagnosis methods face difficulties in detecting sideband waves due to increased spectrum intensities near the power supply frequency when load torque variations occur in induction motors, making it hard to identify abnormalities.

Innovation Solution

A load abnormality detection device that includes a current input unit, FFT analysis unit, peak detection calculation unit, averaging calculation unit, sideband wave extraction unit, and spectrum peak series determination unit to detect abnormalities based on a power spectrum series of sideband waves, ensuring peak detection and noise reduction through averaging.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If conventional facility abnormality diagnosis method is used, then sideband wave detection is performed, but detection accuracy deteriorates when load torque varies due to increased spectrum intensities near power supply frequency

Engineering Contradiction:
Improvesideband wave detection accuracyVSAvoiddetection reliability under varying load torque
Core Design Contradiction:
Measurement precisionVSReliability

Solution Approach 1:

The patent applies preliminary action by performing multiple FFT analyses on current waveforms captured during stable operating periods before the actual abnormality detection. By pre-processing the current waveform data and storing the processed results, the system prepares clean reference data that can be reliably compared during abnormality detection, even when load torque varies during operation.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The patent uses an intermediary approach by introducing a comparison mechanism between the processed current waveform data and reference values. The system compares the processed current waveform with reference current waveform data to detect abnormalities, using this intermediary comparison step to filter out noise and variations caused by load torque changes, thereby improving detection accuracy.

Inventive Principle:
Principle #24Intermediary (Mediator)

2Quantity of substance

If spectrum analysis is performed on varying load conditions, then more operational data is captured, but noise interference increases making peak detection difficult

Engineering Contradiction:
Improveoperational data capturedVSAvoidnoise interference
Core Design Contradiction:
Quantity of substanceVSObject-affected harmful factors

Solution Approach 1:

The system performs preliminary FFT analysis on current waveforms during stable operation periods before abnormality detection. By pre-processing the data and storing the results, it accumulates clean operational data that can be reliably used for comparison, increasing the quantity of useful operational data while maintaining signal quality.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The patent extracts only the relevant stable operation periods from the overall operational data for analysis. By selecting and processing only the current waveforms captured during stable operation, it separates the useful operational data from the noisy varying load conditions, thereby increasing the quantity of reliable data while reducing noise interference.

Inventive Principle:
Principle #2Taking out (Extraction)

3Object-affected harmful factors

If multiple power spectrums are averaged, then signal intensity of noise peaks is reduced, but sideband wave extraction becomes more critical

Engineering Contradiction:
Improvenoise signal intensityVSAvoidsideband wave extraction difficulty
Core Design Contradiction:
Object-affected harmful factorsVSDifficulty of detecting and measuring

Solution Approach 1:

The system performs preliminary FFT analysis and averaging during stable operation periods before abnormality detection. By pre-computing the average of multiple power spectrums during normal operation, it reduces noise signal intensity in advance, making the subsequent sideband wave extraction and abnormality detection more reliable even when load torque varies.

Inventive Principle:
Principle #10Preliminary action

Data Source

PatentEP3474029B1Load abnormality detection device
Publication Date: 2022.03.30 MITSUBISHI ELECTRIC CORP
  • EP3474029B1 patent drawingFigure 1
  • EP3474029B1 patent drawingFigure 2
  • EP3474029B1 patent drawingFigure 3~4

AI summary

To provide, even in the case of an electric motor of which load torque varies, a load abnormality detection device that can detect an abnormality in a load driven by the electric motor on the basis of a spectrum peak series of a sideband wave that occurs in a peak shape at both sides in the vicinity of a power supply frequency. In the load abnormality detection device, the current of an electric motor 5 is detected by a current detector 4 and inputted from a current input unit 10. Then, in a logic calculation unit 11, a plurality of power spectrum analysis results each obtained by frequency analysis being performed on the current waveform when the current is stable are subjected to an averaging process. Then, a sideband wave is detected from a power spectrum analysis result obtained through the averaging process, and the presence/absence of an abnormality in a load is determined on the basis of a spectrum peak series. When occurrence of an abnormality is determined, an alarm output unit 21 outputs an alarm.