Padé Approximant Neural Networks for Enhanced Electric Motor Fault Diagnosis Using Vibration and Acoustic Data

2025-07-03

AI Summary

To address the limited modeling capability of multi-source vibration and acoustic signals in induction motor fault diagnosis, this paper proposes Padé Approximation Neural Networks (PadéNets), the first deep learning architecture to explicitly embed Padé approximation theory. This integration significantly enhances nonlinear representational capacity while remaining compatible with unbounded activation functions such as Leaky ReLU. Evaluated on the publicly available University of Ottawa dataset, PadéNets achieve fault classification accuracies of 99.96%, 98.26%, 97.61%, and 98.33% across four sensor channels—consistently outperforming baseline models including 1D CNNs and Self-Organizing Neural Networks (Self-ONNs). The results empirically validate the efficacy of Padé approximation as a structural inductive bias in neural network design. Moreover, this work establishes a novel multimodal fault diagnosis paradigm for rotating machinery, advancing the integration of classical approximation theory with deep learning for industrial condition monitoring.

Technology Category

Knowledge Representation and Reasoning: Diagnosis and Abductive ReasoningMachine Learning: Multimodal LearningNatural Language Processing: Learning & Optimization for NLP

Application Category

Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphsSystems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applicationsResponsible Web: Machine-in-the-loop, human agency and autonomy

Abstract

Purpose: The primary aim of this study is to enhance fault diagnosis in induction machines by leveraging the Padé Approximant Neuron (PAON) model. While accelerometers and microphones are standard in motor condition monitoring, deep learning models with nonlinear neuron architectures offer promising improvements in diagnostic performance. This research addresses the question: Can Padé Approximant Neural Networks (PadéNets) outperform conventional Convolutional Neural Networks (CNNs) and Self-Organized Operational Neural Networks (Self-ONNs) in diagnosing electrical and mechanical faults using vibration and acoustic data? Methods: We evaluate and compare the diagnostic capabilities of three deep learning architectures: one-dimensional CNNs, Self-ONNs, and PadéNets. These models are tested on the University of Ottawa's publicly available constant-speed induction motor datasets, which include both vibration and acoustic sensor data. The PadéNet model is designed to introduce enhanced nonlinearity and is compatible with unbounded activation functions such as Leaky ReLU. Results and Conclusion: PadéNets consistently outperformed the baseline models, achieving diagnostic accuracies of 99.96%, 98.26%, 97.61%, and 98.33% for accelerometers 1, 2, 3, and the acoustic sensor, respectively. The enhanced nonlinearity of PadéNets, together with their compatibility with unbounded activation functions, significantly improves fault diagnosis performance in induction motor condition monitoring.

Problem

• Enhancing fault diagnosis in induction machines using PAON model
• Comparing PadéNets with CNNs and Self-ONNs for motor faults
• Improving diagnostic accuracy with vibration and acoustic data

Innovation

• PadéNets enhance nonlinearity in fault diagnosis
• Uses vibration and acoustic sensor data
• Compatible with unbounded activation functions