Jul 2026· Journal of Visualized Experiments· Vol 233· 0 citations
Medicine
Abstract
As predictive maintenance transitions from the data-centric paradigm of Industry 4.0 to the sustainable, human-centric framework of Industry 5.0, diagnosing servo motor conditions faces the dual challenges of data scarcity and a profound lack of labeled fault samples. To address this cold-start problem, we present a pseudo-supervised machine learning framework evaluated on a custom five-channel dataset comprising 199 servo motor telemetry samples (current, voltage, temperature, humidity, and vibration). The methodology integrates hard structural partitioning (k-means) and soft posterior confidence estimation (Gaussian Mixture Models) to characterize operating modes without prior annotation. Concurrently, an Isolation Forest model quantifies anomaly intensity and establishes a dynamic quantile-based threshold. A critical innovation of this research is the deterministic risk mapping derived from engineering priors; it defines the "high-risk" (abnormal) state by inversely weighting the physical safety margins of the sensors. This mechanism strictly maps unsupervised clusters to binary pseudo-labels. These labels are subsequently used to supervise downstream discriminators (Random Forest and Support Vector Machine). The final online diagnostic outputs a score-level fusion of the classifier probability and the GMM posterior, gated by the anomaly threshold. Quantitative evaluation demonstrates that the Random Forest model achieved a perfect F1 score of 1.000, while the comparative SVM yielded an F1 score of 0.997, proving the framework to be a robust, interpretable, and highly accurate solution for cold-start industrial health monitoring.
Unplanned failures of induction motors impose serious operational and financial penalties on industrial facilities, yet the fault signatures that precede such failures are detectable well in advance through careful sensor instrumentation and data-driven analysis. This paper presents an end-to-end Internet-of-Things (IoT) predictive maintenance scheme based on two off-the-shelf sensors: a DS18B20 one-wire digital thermometer and 2 piezo vibration sensor modules, with an ESP32 edge device for running the full machine learning pipeline offline, independent from any cloud services. Four operating scenarios are considered: healthy condition, BPFO (bearing outer race fault), misaligned shaft, and rotor imbalance. Based on 1-second sampling intervals, 17 descriptors are derived, including statistics in the time domain, Fourier harmonic peaks, energy ratios between different frequency bands, and temperature gradients measured across all sensors. A dual-stage feature selection method using mutual information (MI) score and Random Forest mean decrease impurity (MDI) ranking reduces the number of features to the 10 most relevant descriptors, reducing the computational complexity by 41% at the expense of 5.6% F1-macro. On a balanced 600-sample synthetic dataset, the resulting Random Forest classifier attains 87.3% hold-out accuracy, 91.0±1.9% five-fold cross-validation accuracy, and a macro area-under-the-ROC-curve of 0.980. End-to-end inference takes just 39 ms on the ESP32, easily meeting the 200 ms requirement for real-time alerting.
Akash Mastud, Dhiraj Vaidya, Azaroddin Sayyed et al.· International Conference on...· 0 citations
Electric motors are the backbone of industrial productivity, safety, and energy efficiency, making their reliable operation essential to modern manufacturing. As Industry 4.0 reshapes how equipment is managed, predictive maintenance has emerged as a critical capability. Yet traditional condition monitoring still depends largely on routine servicing, periodic inspections, centralised processing, and predefined thresholds – approaches that limit early fault detection and offer little support for real-time monitoring. This research proposes an AI-driven, edge-based system for electric motor health monitoring and predictive maintenance using multi-sensor data. The system integrates vibration, temperature, and electrical current sensors for continuous data acquisition, and a deep learning classifier was developed to predict motor health and detect faults at an early stage. To achieve low latency, reduced bandwidth usage, and fast on-site decision-making, the model was optimised and deployed using TinyML, an edge AI framework. A companion web application supports remote monitoring, real-time visualisation, and maintenance alerts. Using an industrial dataset of 8,000 records, the TinyML model was trained, optimised, and deployed on an Espressif ESP32 microcontroller. Through spectral analysis and 8-bit quantisation, the system achieved a classification accuracy of 92.3% with an inference latency of just 16 ms, while consuming only 13.4 KB of RAM and 95 KB of Flash memory. These results demonstrate that high-speed, decentralised predictive maintenance is feasible on low-power hardware, offering a scalable solution for the energy and manufacturing sectors, reducing unplanned downtime and enabling rapid response to abnormal operating conditions.
F. Haruna, M. Abdulraheem, I. O. Durotoye et al.· SPE Nigeria Annual Internati...· 0 citations
Air-operated double-diaphragm (AODD) pumps in industrial sludge transfer suffer from gradual performance degradation due to rheological variations and component wear, yet conventional monitoring relies on scarce labeled fault data. This paper presents an unsupervised multi-sensor framework that requires no fault labels, integrating physics-informed dual health indices, HI-P (sludge load) and HI-V (mechanical stress), with a Gaussian Mixture Model anomaly detector and a physics residual attribution module. Governing equations motivate the use of these indices from five sensors: inlet and outlet flow meters (100 Hz), an air pressure transducer (100 Hz), and inlet and outlet accelerometers (1652 Hz). Trained on one healthy baseline day (86,218 one-second windows), the Gaussian Mixture Model achieves 100% day-level classification performance on the evaluated dataset (F1 = 1.00) across 455,201 test windows from nine operating days, with window-level receiver operating characteristic area under the curve (ROC-AUC) = 0.8580 and precision–recall AUC (PR-AUC) = 0.9082. Residual attribution analytically confirms that pressure residuals drive Episode 1 (HI-P peak 3.63 times baseline, Cohen’s d = 1.70) and vibration residuals drive Episode 2 (HI-V peak 5.44 times the baseline, d = 4.10), providing empirical support for the proposed physics-informed formulation without requiring fault labels. Comparisons with four unsupervised benchmarks confirm that this is the only approach that simultaneously enables label-free operation, physics-driven features, exact attribution, real-world deployment, and perfect day-level F1.
Seong-Wook Kim, A. B. Kareem, J. Hur· Italian National Conference...· 0 citations
In Industry 4.0 manufacturing, predictive maintenance has emerged as a key focus to ensure the reliability of operations, minimize downtime, and contribute to equipment safety. This study proposes a predictive maintenance framework for additive manufacturing systems (AMSs) using an explainable artificial intelligence (XAI) and machine learning (ML) approach, which is supported by real-time multi-sensor monitoring and is synchronized with a digital twin architecture. A heterogeneous dataset of over 10,000 observations and 13 sensor attributes was gathered from a sensing architecture with ESP32 cameras designed to handle heterogeneous signals from thermal, environmental, mechanical and safety-related sensors. Domain-aware feature engineering was done to gain insights into operation indicators such as temperature instability, vibration degradation, smoke risk, humidity anomalies and aggregated maintenance risk scores. Multiple predictive models, such as the Random Forest model, XGBoost model, Logistic Regression model, K-Nearest Neighbours classifier, Multi-Layer Perceptron model, and ARIMA forecast model, were comparatively assessed under highly imbalanced maintenance conditions. The results showed that ensemble learning methods, especially XGBoost and MLP, had better recall and ROC-AUC for fault detection of maintenance-critical problems. SHAP and LIME analyses then showed that a number of physically meaningful indicators, including thermal instability, vibration anomalies, smoke-related features, and aggregated risk scores, have a significant impact on maintenance predictions and hence provide better operational interpretability and maintenance transparency.
Chitranjanjit Kaur, S. Chopra, Chitta Ranjan Tripathy· Automation· 0 citations
The integration of Internet of Things devices in modern industrial environments has fundamentally transformed paradigms of machine health monitoring and predictive maintenance. However, industrial environments are inherently dynamic, leading to continuous fluctuations in operational conditions such as varying loads, speeds, and environmental temperatures. These fluctuations induce significant nonstationarity in the collected time-series data, rendering conventional assumption of independent and identically distributed data invalid and severely degrading the performance of standard deep learning models for fault diagnosis. This paper presents a comprehensive theoretical and methodological framework for nonstationarity-aware time-series modeling specifically tailored for industrial fault detection and diagnosis. We propose an advanced neural network architecture that dynamically adapts to distribution shifts through a continuous domain adaptation mechanism and an adaptive attention module. By incorporating localized statistical normalization and non-stationary feature alignment, the proposed approach effectively mitigates the adverse effects of concept drift without requiring explicit operational condition labels. Extensive theoretical analysis is provided to formalize the bounds of distribution divergence in degrading machinery. Furthermore, rigorous empirical evaluations on multiple complex industrial datasets demonstrate that the proposed framework significantly outperforms existing state-of-the-art diagnostic models in terms of accuracy, robustness, and generalization capabilities under severe operational transitions. The findings emphasize the critical necessity of embedding nonstationarity awareness directly into the optimization objectives of predictive models, paving the way for more resilient and autonomous industrial health management systems.
Emily A. Young, Hannah Turner· International journal of inf...· 0 citations
This study provides a systematic robustness evaluation of classical machine learning for vibration-based bearing fault detection in low-RPM internal combustion engines (1000–2000 RPM) across a controlled temperature × humidity grid, a regime underrepresented in benchmark datasets that emphasise high-speed applications. A publicly available dataset from a 658cc engine (–10 °C–45 °C, 0%–100% humidity) was analysed; vibration features were derived from the non-zero channels of a tri-axial acquisition, with the bearing-housing vibration carried primarily by channel Ch3. To prevent temporal leakage, 390 263 continuous measurements were aggregated into 89 steady-state units, each spanning 90 s, yielding a deliberately independence-preserving but low sample-to-feature ratio (89:92). Four algorithms Random Forest, Support Vector Machine, Logistic Regression, and Neural Network were evaluated using stratified 5-fold cross-validation. All models achieved apparent accuracy exceeding 95%, with Random Forest performing best (97.8% ± 2.7%), but no statistically significant differences were found (Friedman test, p = 0.732). Vibration features, particularly crest factor and root mean square, provided the greatest discriminative power, while environmental factors accounted for less than 17% combined importance. The near-perfect linear separability (98.2% with Logistic Regression) indicates that the dataset’s binary, controlled-laboratory labelling rather than intrinsic bearing-degradation physics drives the clean classification. Accordingly, the reported accuracies are apparent upper-bound estimates from an exploratory study, not expected field performance; validation on 500–1000 or more samples with progressive-degradation labelling is essential before any operational claim can be supported.
P. Pugazhendi, Vinoth Vishwanathan, Aadil Arshad Ferhath et al.· Engineering Research Express· 0 citations