Work

Low-Complexity Vibration-Spectrum Feature Learning for Early-Stage Inter-Turn Short-Circuit Diagnosis in Three-Phase Induction Motors

Year: 2026

Type: conference-paper

Abstract: Three-phase induction motors (TIM) are widely used in industrial applications, and in-ter-turn short-circuit faults represent one of the most critical incipient failure modes. This paper proposes a lightweight and interpretable vibration-based fault diagnosis framework that combines frequency-domain feature extraction with a multilayer perceptron (MLP) classifier. Vibration signals acquired by MEMS (Micro-Electro-Mechanical Systems) accelerometers were segmented and processed using the Fast Fourier Transform (FFT). From the resulting spectra, a compact set of statistical features—including energy, spectral centroid, bandwidth, kurtosis, and skewness—was extracted and used as input to the MLP. The method was evaluated under healthy conditions and multiple inter-turn short-circuit fault scenarios, considering different affected phases and severity levels. Experimental results demonstrate competitive classification performance with significantly reduced computational cost compared to deep learning approaches.

Source: International Electronic Conference on Machines and Applications (IECMA)

DOI: 10.3390/engproc2026145014

Authors: Bruno da Silva Nassula, Guilherme Beraldi Lucas, André Luiz Andreoli

Institution: Universidade Estadual Paulista (Unesp)

Language: English

Topic: Machine Fault Diagnosis Techniques

Subfield: Control and Systems Engineering

Field: Engineering

Domain: Physical Sciences

SDG: Industry, innovation and infrastructure

Open Access status: hybrid

Funder: Coordenação de Aperfeiçoamento de Pessoal de Nível Superior