Online Unsupervised Condition Monitoring and Fault Detection for Machine-Level Motors Using Generic Feature Learning

Condition monitoring and fault detection of machine-level motors are critical in industrial environments. Traditional supervised approaches, which rely on extensive labeled fault data, are often impractical when such data are limited or unavailable. Moreover, existing methods face challenges in adapting to dynamic operating conditions and distinguishing subtle faults from normal variations. Additionally, the construction of real-time statistical indicators for adaptive monitoring still has some drawbacks. To address these challenges, this article proposes a novel online unsupervised condition monitoring framework for machine-level motors. In the offline phase, a Siamese network-based feature extractor is utilized to learn robust feature representations solely from normal operational data whose noise is introduced by machine operation. In the online phase, an orthogonal low-rank transformation (OLT) is introduced to enhance the separability of normal and faulty features, while a regularization term ensures that the learned features approximate Gaussian, thereby improving model generalization in evolving operational conditions. A dynamic monitoring strategy is then developed to construct real-time statistical indicators based on the learned Gaussian features. These indicators are continuously updated as new data is received, enabling adaptive, real-time fault detection without the need for labeled fault data. Experimental analysis on two motor datasets and comparative studies demonstrates the effectiveness of the proposed method, which can contribute to the development of efficient, data-driven monitoring strategies for practical deployment in real-world industrial applications.

• Publication year

  2025

• Venue

  IEEE Transactions on Instrumentation and Measurement

• Publication date

  Unknown publication date

• Fields of study

  Not labeled

• Identifiers
  DOI 10.1109/TIM.2025.3628436
• External record

  Open on Semantic Scholar

• Source metadata

  Semantic Scholar

• No claims are published for this paper.

• No concepts are published for this paper.

• A Real-Time Adaptive Fault Diagnosis Scheme for Dynamic Systems With Performance Degradation

  2024cited by this paper
• A systematic overview of health indicator construction methods for rotating machinery

  2024cited by this paper
• A new unsupervised health index estimation method for bearings early fault detection based on Gaussian mixture model

  2024cited by this paper
• Condition Monitoring using Machine Learning: A Review of Theory, Applications, and Recent Advances

  2023cited by this paper
• A Novel Local Binary Temporal Convolutional Neural Network for Bearing Fault Diagnosis

  2023influential reference
• Unsupervised Deep Tensor Multitask Anomaly Detection With Rule Adaptation for Online Early Fault Evaluation

  2023cited by this paper
• Unsupervised Deep Multitask Anomaly Detection With Robust Alarm Strategy for Online Evaluation of Bearing Early Fault Occurrence

  2022cited by this paper
• Self-Supervised Simple Siamese Framework for Fault Diagnosis of Rotating Machinery With Unlabeled Samples

  2022influential reference
• Latent adversarial regularized autoencoder for high-dimensional probabilistic time series prediction

  2022cited by this paper
• An unsupervised feature learning based health indicator construction method for performance assessment of machines

  2022cited by this paper
• A survey of machine-learning techniques for condition monitoring and predictive maintenance of bearings in grinding machines

  2021cited by this paper
• Real-time monitoring and diagnosis scheme for IoT-enabled devices using multivariate SPC techniques

  2021cited by this paper
• Self-starting process monitoring based on transfer learning

  2021cited by this paper
• Deep Variational Autoencoder Classifier for Intelligent Fault Diagnosis Adaptive to Unseen Fault Categories

  2021cited by this paper
• A review on deep learning in machining and tool monitoring: methods, opportunities, and challenges

  2021cited by this paper
• A novel temporal convolutional network with residual self-attention mechanism for remaining useful life prediction of rolling bearings

  2021influential reference
• Siamese Neural Network-Based Supervised Slow Feature Extraction for Soft Sensor Application

  2021cited by this paper
• Deep Residual Shrinkage Networks for Fault Diagnosis

  2020cited by this paper
• A Hybrid Prognostics Approach for Estimating Remaining Useful Life of Rolling Element Bearings

  2020cited by this paper
• A review of machine learning kernel methods in statistical process monitoring

  2020cited by this paper
• Signal based condition monitoring techniques for fault detection and diagnosis of induction motors: A state-of-the-art review

  2020cited by this paper
• Review on Machine Learning Algorithm Based Fault Detection in Induction Motors

  2020cited by this paper
• Nonparametric Phase-II control charts for monitoring high-dimensional processes with unknown parameters

  2020cited by this paper
• Graph-Based Change Detection for Condition Monitoring of Rotating Machines: Techniques for Graph Similarity

  2019cited by this paper
• Toward a better monitoring statistic for profile monitoring via variational autoencoders

  2019cited by this paper
• Machine Learning-Based Fault Diagnosis for Single- and Multi-Faults in Induction Motors Using Measured Stator Currents and Vibration Signals

  2019cited by this paper
• Sensor Fusion via Statistical Hypothesis Testing for Prognosis and Degradation Analysis

  2019cited by this paper
• Unsupervised electric motor fault detection by using deep autoencoders

  2019cited by this paper
• Early Fault Detection Approach With Deep Architectures

  2018cited by this paper
• Review and big data perspectives on robust data mining approaches for industrial process modeling with outliers and missing data

  2018cited by this paper
• Performance-Based Gas Turbine Health Monitoring, Diagnostics, and Prognostics: A Survey

  2018cited by this paper
• PCA in High Dimensions: An Orientation

  2018cited by this paper
• OLE: Orthogonal Low-rank Embedding, A Plug and Play Geometric Loss for Deep Learning

  2017cited by this paper
• Industrial maintenance decision-making: A systematic literature review

  2017cited by this paper
• A convolutional neural network based feature learning and fault diagnosis method for the condition monitoring of gearbox

  2017cited by this paper
• Sensor Fault Diagnosis for Structural Health Monitoring Based on Statistical Hypothesis Test and Missing Variable Approach

  2017cited by this paper
• Health Index-Based Prognostics for Remaining Useful Life Predictions in Electrical Machines

  2016cited by this paper
• Statistical Learning Methods Applied to Process Monitoring: An Overview and Perspective

  2016cited by this paper
• Fuzzy exponentially weighted moving average control chart for univariate data with a real case application

  2014cited by this paper
• Performance of t control charts in short runs with unknown shift sizes

  2013cited by this paper
• A Regularized Hotelling’s T2 Test for Pathway Analysis in Proteomic Studies

  2011cited by this paper
• Prognostics in Battery Health Management

  2008cited by this paper
• IN SHANGHAI, CHINA

  2007cited by this paper
• Convex analysis approach to d.c. programming: Theory, Algorithm and Applications

  1997cited by this paper
• Characterization of the subdifferential of some matrix norms

  1992cited by this paper

• No citing papers are available for this paper.