Search Results (121)

Reliable motor-current-based bearing diagnosis requires evaluation on unseen physical bearings and operating conditions. This study uses the Paderborn University benchmark, acquired from a 425 W permanent-magnet synchronous motor (PMSM) test rig, to evaluate time–frequency deep transfer learning under strict bearing-level grouping. Four representations—continuous wavelet transform (CWT), short-time Fourier transform (STFT), wavelet synchrosqueezed transform (WSST), and Fourier synchrosqueezed transform (FSST)—are combined with pretrained CNN backbones across four binary targets: aged-only A/B and artificial-plus-aged C/D, with mixed-fault bearings excluded/included within each pair. The workflow includes pooled-condition candidate discovery, exploratory Main-split leave-one-operating-condition-out (LOCO) screening, and a retrospective multi-split LOCO audit. The audit contains 288 crossed condition–split–seed evaluations. Because pooled test summaries and Main-split LOCO results informed later stages, these evaluations provide descriptive robustness evidence rather than an independent post-selection test. Target A achieved the highest all-split mean balanced accuracy (0.736 for CWT–EfficientNetB0). The pairs for Targets B and C were near-ties, and the Target D ordering reversed when Main was excluded. Across the eight audited candidates, mean sensitivity ranged from 0.618 to 0.948, whereas specificity ranged from 0.092 to 0.564. Target D combined approximately 0.89 sensitivity with an approximately 0.90 false-alarm rate. Thus, operating condition, fault-class composition, bearing split, and error-cost priorities all affect model interpretation. A matched current-domain baseline audit added 336 evaluations using handcrafted-feature RBF–SVM and Random Forest models and a compact raw-current 1D-CNN. The results show that instability is broader than the TF–CNN pipeline but is not uniform across model families: TF candidates were clearly stronger for Targets A and C, feature-based models were stronger for Target B, and Target D remained mixed and protocol-sensitive. A complementary bearing-level source-group analysis quantified six healthy/fault-source categories across the 37 current features; among the eight features with the largest mean between-group variance fraction, only spectral entropy and dominant power fraction preserved the same mixed-versus-non-mixed contrast direction across all four operating conditions. An additional matched Target D reference held the binary target, physical-bearing split, held-out condition, seed, fourth-order FSST representation, ResNet-50 backbone, and training settings fixed while changing the sensing channel. Across 24 matched runs, vibration showed descriptively higher mean balanced accuracy (0.642 vs. 0.503) and specificity (0.476 vs. 0.146), while sensitivity was slightly lower (0.807 vs. 0.859); the comparison does not establish universal modality superiority. Pooled-condition performance is useful for candidate discovery, but credible condition-generalization claims require explicit separation of exploratory selection and confirmatory testing. The numerical findings are specific to the evaluated PMSM benchmark and do not establish universal performance across electric-machine types. Full article

This paper addresses fault diagnosis in nonlinear systems where reliable mathematical models are unavailable and only input–output measurements are accessible. The proposed methodology consists of three stages. First, a data-driven identification stage is performed using an Adaptive Neuro-Fuzzy Inference System (ANFIS) to capture the nonlinear dynamics of the system from fault-free sensor data. This procedure yields a set of convex Takagi–Sugeno (TS) models representing the system dynamics. In the second stage, fault detection is achieved using zonotopic proportional–integral (PI) observers with convex structures. Robustness against parametric uncertainty and sensor noise is ensured through an ${\mathcal{H}}_{\infty }$ formulation expressed as a set of linear matrix inequalities (LMIs). Finally, fault isolation is carried out using a fault signature matrix (FSM). The zonotopic framework provides adaptive set-based residual bounds that act as adaptive thresholds for fault detection, while structured residual activation patterns enable reliable fault isolation. The proposed approach is evaluated on a continuous stirred tank reactor (CSTR) under sensor faults and incipient process faults in the presence of measurement noise and compared with representative data-driven methods. Results demonstrate improved diagnostic accuracy and reduced false-alarm rates while maintaining timely fault detection and reliable isolation. Full article

The stability and reliability of temperature sensors in industrial cooling systems are critical to process quality, energy efficiency, and operational safety. However, existing approaches lack systematic stability metrics and intelligent predictive capabilities. This study proposes an AI-driven framework integrating stability feature engineering with machine learning models for fault identification and early prediction of temperature sensors in power plant cooling systems. The framework introduces three physics-based stability indicators—rolling standard deviation (σ_roll), variation intensity index (VII), and short-term variation magnitude (ΔT_short)—to quantify sensor signal quality. These features, combined with operational parameters, are used to train support vector machine (SVM) and Long Short-Term Memory (LSTM) models for binary classification. The framework is validated using over 260,000 one-minute records per unit collected from three parallel steam-turbine generating units (Units 1, 2, and 3) of the same coastal thermal power plant. Each unit is served by an independent once-through seawater cooling loop instrumented with redundant Pt-100 temperature sensors at the inlet and outlet manifolds; the three units differ in their operating profile—Unit 1 operates under variable load with frequent cold-start events, Unit 2 under moderate variable load, and Unit 3 under stable high-load conditions—with data collected at 1 min intervals from January to June 2025. Under an explicitly anomaly-positive evaluation, with the full confusion matrix reported for every unit and model, classification performance is limited and strongly unit-dependent. In real-time identification, AUC-based ranking ability varies across units (SVM AUC = 0.65, 0.75, and 0.98 for Units 1–3; LSTM AUC = 0.66, 0.31, and 0.52), but under the extreme class imbalance (anomaly rate ≈ 0.07–0.13% in the test partitions), the calibrated operating-point precision and F1-scores remain low for all unit–model combinations (F1 ≤ 0.26, MCC ≤ 0.28). McNemar’s test indicates statistically significant paired differences for Units 1 and 2 but not for Unit 3. These results show that, on this dataset, neither model attains reliable anomaly classification, and that all reported metrics must be interpreted together with the disclosed confusion-matrix counts and severe class imbalance. The primary contribution of the framework is therefore methodological—physics-based stability indicators, redundant sensor cross-checking, and an operational false-alarm analysis—rather than high-accuracy prediction, and the study highlights the difficulty of learning-based prediction for rare, rule-defined thermal sensor anomalies. Full article

Strong broadband noise can mask the weak impact responses produced by bearing faults, while wavelet reconstructions selected only by signal-domain scores may not provide useful inputs for diagnosis. To address this problem, this paper presents WPMSR-1D-ResNet, a training-aware wavelet-domain controlled-noise augmentation framework. PKPM-guided particle swarm optimization first generates scale-specific perturbation candidates in the DWT detail coefficients. Signal-fidelity constraints remove distorted reconstructions, and a proxy CNN with identity fallback selects a global candidate according to validation Macro-F1. The final 1D-ResNet is trained jointly with the measured waveform and the selected augmented view, whereas inference uses only the raw signal. Under one fixed data construction and a common fixed 20-epoch budget, WPMSR-1D-ResNet achieved $0.8311\pm 0.0607$ Macro-F1 at the predefined CWRU low-SNR endpoint and ranked third among eleven methods, placing it within the leading statistical group. Its paired mean exceeded raw-signal 1D-ResNet and per-slice PKPM replacement by 0.05582 and 0.05227, respectively, with gains in nine of ten paired computational seeds. Mixed-SNR training increased mean Macro-F1 across eight mismatched-noise conditions from $0.5463\pm 0.1320$ to $0.6105\pm 0.1292$. The method ranked second on PU and third on acquisition-held-out AT data; on AT, it reduced the normal-state false-alarm rate from 0.2854 to 0.1646 while maintaining 0.9708 fault sensitivity. Raw-only inference required 0.266 ms per slice. WPMSR-1D-ResNet, therefore, aligns wavelet augmentation with diagnostic performance, preserves useful waveform information, and removes wavelet reconstruction and PSO from the deployment path. Full article

Up/down counters are ubiquitous in the alarm and fault-confirmation logic of electro-mechanical systems. In aircraft, several electro-mechanical modules provide position feedback for flight control; the Linear Variable Differential Transformer (LVDT) is a representative one, converting mechanical displacement into an electrical signal whose reliable monitoring is critical to flight safety. As such counters are deployed in ever more complex systems and more uncertain environments, rising safety requirements render their heuristic tuning unreliable. To address this challenge, this paper proposes a quantitative, reliability-oriented procedure for counter-based monitors, which replaces heuristic parameter tuning. Both healthy and faulty signal distributions are estimated by Kernel Density Estimation (KDE), so the framework handles non-Gaussian noise and FMEA-weighted failure modes. The threshold-and-counter logic is modeled as a finite-horizon absorbing Discrete-Time Markov Chain (DTMC), which yields the false-confirmation probability, missed-detection probability, and detection delay over a bounded horizon instead of long-run rates. Thresholds and counter parameters are then synthesized offline, leaving a lightweight online monitor that needs only threshold comparison and integer counter updates. We evaluate the method on an LVDT sum-voltage monitor using real aircraft healthy measurements and Simulink-based fault injection with 45 detectable modes, assessing the synthesized monitor on $\mathrm{558,001}$ real measured samples and an FMEA-driven fault population. Results show that, among the compared confirmation logics, our workflow yields a counter that meets the ${10}^{-9}$ false-confirmation target and the ${10}^{-6}$ missed-detection target while attaining the lowest detection delay. Full article

Edge monitoring balances diagnostic information against high-rate sensing costs. We formulate active acquisition: Low-cost signals screen each window, while high-information channels are acquired when alarms require confirmation or the screen is uncertain or out of distribution. The confirmation-complete controller includes an optional watchdog to bound blind intervals. On a five-stress benchmark, it retains 98.42% Macro-F1 versus 99.02% for always-full sensing, with 77.59% activation and 19.27% lower normalized resource costs. CWRU evaluation yields 99.25 ± 0.41% Macro-F1 at 83.30% activation. Paderborn condition-transfer evaluation matches always-full sensing at 91.08 ± 17.55% but activates every window; few-shot commissioning increases 15-bearing performance from 58.42 ± 18.60% to 76.52%. When each damage type is excluded from training and bearing identities are separated, the policy activates for 85.25 ± 18.84% of unseen-fault windows versus 83.61 ± 25.69% for a learned error gate. Across 40 healthy-to-unseen replays, 95% activate within five decision windows; an $H=10$ watchdog limits the inactive run to nine. These rates measure escalation, not unknown-class identification. An ESP32 pilot verifies telemetry feasibility but not sensor-rail power or end-to-end diagnosis. Reduced sensing is credible when the confirm stage transfers; otherwise, commissioning or conservative acquisition is required. Full article

When multi-cylinder diesel engines operate under variable speed and load conditions, the distribution of normal vibration responses drifts with the operating conditions. This makes it difficult for diagnostic models to distinguish normal operating-condition fluctuations from genuine fault deviations, leading to false alarms or missed detections. Meanwhile, fault samples are usually limited in practical applications. To address these problems, this study proposes an anomaly detection method based on multi-feature relational modeling and conditional-memory augmentation. The method performs the cycle-wise alignment of multi-point vibration signals according to the firing phase of each cylinder. It integrates local waveform morphology, impact energy, and energy-centroid information in the non-uniform angular domain to construct a raw–relative dual relational representation. It further uses speed conditions to modulate latent features and employs a sparse normal memory to constrain reconstruction sources, enabling the model to learn normal relational patterns under different operating conditions using only normal samples. Tests involving misfire, intake-valve clearance anomaly, and exhaust-valve clearance anomaly were conducted on a TBD234V12 diesel-engine test bench. The proposed method achieved an accuracy, true positive rate (TPR), F1-score, and area under the receiver operating characteristic curve (AUROC) of 97.44%, 98.98%, 98.30%, and 98.88%, respectively, with a false-positive rate (FPR) of 7.21% under the sample-level alarm definition. The results show that the method reduces the interference of operating-condition-induced normal-pattern drift with anomaly determination and improves the accuracy of fault warning within the range of the operating conditions covered in this study. Full article

This paper presents an interval-observer-based actuator fault detection and isolation (FDI) method for underactuated autonomous underwater vehicles (AUVs) under bounded hydrodynamic uncertainty and time-varying ocean currents. A locally frozen linear time-invariant (LTI) representation enables deterministic set-membership analysis, and the robust adaptive propagated interval observer (RAPIO) propagates admissible center–radius state bounds within a Lyapunov framework. Adaptivity is introduced through a reinforcement learning (RL)-augmented uncertainty-bound modulation mechanism, where an offline-trained agent scales a nonnegative channel-wise slack term without modifying the scheduled observer-gain rule or the nominal center predictor. Under the stated observer and disturbance-envelope conditions, positivity, stability, and diagnostic-channel inclusion hold for any bounded learning signal. Actuator loss-of-effectiveness (LoE) faults are represented through the actuator-effectiveness channel and detected through interval-consistency violations, enabling axis-wise isolation of surge, yaw-rate, and pitch-rate actuator faults. The same schedule-blind decision layer is additionally evaluated with structurally distinct additive-bias and stuck/jam actuator models. All stuck/jam events are detected, and bias-magnitude sweeps identify channel-wise 100%-detection boundaries with zero false alarms. A structured 72-case scenario sweep shows reliable detection, strong false-alarm rejection, and acceptable detection delays compared with benchmark observers. Full article

Industrial process safety systems are predominantly reactive: alarms activate after limits are crossed, faults are diagnosed after deviations develop, and HAZOP knowledge remains offline during operation. This paper proposes Predictive Operational Safety Engineering (POSE) as an emerging research paradigm in which operational safety is treated as a continuously forecastable state rather than a post-event classification, shifting the operational question from what has gone wrong? to how much safe operating time remains, and which intervention is most urgent? Four integrated predictive safety metrics anchor the framework: Remaining Safety Margin (RSM), quantifying the normalized distance between the predicted process trajectory and the nearest safety boundary; Remaining Safe Operating Time (RSOT), estimating when that boundary will be crossed under the current trajectory; the Operational Vulnerability Index (OVI), combining margin depletion rate, safeguard availability, and consequence severity into a single intervention-urgency signal; and Predictive Safety Confidence (PSC), the probability that a specific named operator intervention can be executed to completion before the predicted safety boundary is crossed, coupling prediction uncertainty with action execution time. The Predictive Operational Safety Twin (POST) is proposed as a three-layer reference architecture implementing POSE through predictive process intelligence, predictive safety intelligence, and human safety intelligence. The paper synthesizes six research streams, positions POSE against seven adjacent disciplines, states ten guiding principles, and formulates a research agenda. As a conceptual narrative review, the paper does not claim empirical validation of POSE. Instead, it establishes the foundational vocabulary, reference architecture, and research agenda required to advance predictive operational safety from an emerging concept toward benchmarked and industrially validated practice. This article constitutes the conceptual and evidence-synthesis phase of a staged research program; subsequent work must test the proposed constructs through benchmark simulation, uncertainty calibration, baseline comparison, operator studies, and industrial case studies. Full article

This study presents an integrated design-to-operation framework for a PV–battery solar highway lighting system supported by IoT-based monitoring for highway-scale deployment. The proposed framework combines pole-level PV-battery sizing, annual energy-reliability simulation, road-level techno-economic optimization, and IoT-based digital monitoring. The Shoubra–Banha Freeway in Egypt, a 40 km corridor with a two-sided lighting arrangement, was selected as the case study. In the pole-level phase, dimming strategies, PV capacities, battery sizes, and battery technologies were evaluated using sequential parametric analysis under a reliability constraint of Loss of Load Probability (LLP) below 1%. The S2 aggressive dimming profile achieved the best operating performance, with an LLP of 0.003, energy reliability of 99.70%, and 2.89 kWh annual unmet load. The minimum feasible PV capacity was 0.8 kW, while the smallest acceptable storage capacity was 4.8 kWh nominal capacity, corresponding to approximately 3.84 kWh usable capacity under an 80% allowable depth of discharge. Among the tested battery technologies, LiFePO₄ achieved the best reliability performance, with an LLP of 0.007 and a 99.25% battery deficit coverage ratio. In the road-level phase, three deployment configurations were compared. Case B, using 12 m pole height and 36 m spacing, was selected as the best-balanced solution, requiring 2224 poles, achieving 99.30% energy reliability, an LLP of 0.007, annual PV generation of 3,178,341 kWh, and total CAPEX of approximately 2.88 × 10⁸ EGP, equivalent to about 5.76 million USD based on an assumed exchange rate of 1 USD = 50 EGP. Lastly, the development of an IoT-based monitoring system design utilizing sector gateways, telemetry variables, alarm conditions, MQTT protocols, and dashboard displays was carried out. The scenario for gateways at 5 km intervals was advised due to its better fault isolation, lower gateway workload, and scalability. This indicates that the suggested approach offers a viable, cost-efficient, and technologically enabled solution for automated solar-powered street lighting systems. Full article

Monitoring the health of bearings in industrial rotating machines is a major challenge for ensuring the reliability and continuous operation of installations. Conventional fault detection methods, based on multivariate control charts such as Hotelling’s T², multivariate exponentially weighted moving average, or multivariate cumulative sum control chart, are limited by the complex nonlinear dynamics of the system. In this article, we propose an innovative monitoring approach based on the Koopman operator, allowing the linearization of a nonlinear system in an observed space and the application of drift detection techniques via an extended T² control chart. The study is based on two experimental approaches: one using controlled simulated data to analyze the responsiveness and robustness of the model, and the other applied to real data from an industrial turbogenerator monitoring the vibrations, temperatures, and speeds of the front and rear bearings. Comparative results show that the Koopman-based T² map detects defects earlier, with better accuracy under noise and a reduced false alarm rate compared to conventional methods. The integration of wavelet preprocessing, statistical feature extraction by sliding windows, and PCA representation of the trajectories enhances the robustness and interpretability of the model. Full article

Following the occurrence of a substation fault, alarm systems frequently exhibit issues such as high-density redundancy, spurious alerts, undetected alarms, and temporal inconsistencies. In addition, the effective integration of heterogeneous multi-source data—including fault waveform records and phasor measurement unit (PMU) streams—remains a nontrivial challenge, further impeding the accuracy of fault diagnosis. To address these limitations, this paper proposes an enhanced framework for multi-source fault event modeling and candidate scenario generation based on an extended temporal constraint network. First, protection operations, circuit breaker status transitions, transient waveform features, and PMU-derived dynamic parameters are abstracted as fault event nodes. On this basis, an event representation model is constructed, incorporating event type, associated equipment, temporal attributes, data source identifiers, and reliability coefficients. Second, the conventional temporal constraint network is refined by jointly embedding temporal constraints, distance constraints, and event confidence metrics, thereby enabling the characterization of uncertain causal-temporal dependencies between fault origins and multi-source observations. Third, a dynamic time-window mechanism is introduced to perform online event clustering, followed by reverse temporal reasoning to generate a set of candidate fault scenarios. These scenarios are subsequently verified, merged, and ranked with reference to waveform and PMU data, facilitating the identification of false alarms, missing alarms, and timestamp anomalies. Finally, case studies conducted on a 220 kV substation and its interconnected transmission network demonstrate that the proposed method achieves a correct scenario coverage rate of 96.8%, a candidate scenario compression rate of 68.7%, and an abnormal alarm recognition accuracy of 89.5%, while sustaining an average scenario generation time of 0.28 s. Full article

Real-time industrial anomaly detection is increasingly shifting from cloud-based diagnosis to edge intelligence deployed close to machines. However, practical industrial scenarios are constrained by scarce fault samples, unknown anomaly types, cross-machine distribution shifts, strict false alarm requirements, data privacy restrictions, and limited edge device resources. To address these challenges, this paper proposes EPC-TinyAD, an energy- and privacy-aware compressed TinyML framework for reliable industrial anomaly detection on resource-constrained edge devices. EPC-TinyAD follows a normal-only learning paradigm and employs a tiny depthwise-separable CNN autoencoder as the deployable student model, guided by a wider teacher autoencoder during training. Instead of relying solely on reconstruction error, the proposed anomaly score integrates spectrogram reconstruction deviation, compact normal-center distance, and teacher–student distillation discrepancy. Masked spectrogram modeling is introduced to enhance few-shot normal representation learning, while domain-adversarial invariant embedding improves cross-machine generalization. To support reliable deployment, split and adaptive conformal thresholding calibrate anomaly decisions under target false alarm rates. Furthermore, federated training with clipped and noisy updates reduces raw industrial data exposure, and energy-aware compression integrates pruning, INT8 size estimation, model export, latency benchmarking, and Pareto analysis. Experiments on industrial anomaly detection data demonstrate that EPC-TinyAD achieves 96.5% accuracy, 95.4% recall, 96.1% F1 score, 0.964 AUROC, and 0.952 AUPRC over five random seeds. These results indicate that EPC-TinyAD provides a reliable, lightweight, privacy-aware, and deployment-oriented framework for industrial edge anomaly detection, while future work will further validate its runtime memory, latency, and power consumption on physical Raspberry Pi-, Jetson-, or MCU-class edge devices. Full article

Aero-engine safety and prognostics and health management (PHM) rely on robust vibration-based fault diagnosis. However, many deep learning studies on rotating machinery are evaluated under random train–test splits that mix hardware instances and may obscure the domain shift faced in deployment. This paper presents a protocol-driven end-to-end baseline for multi-sensor aero-engine-relevant vibration diagnosis on the HIT inter-shaft bearing benchmark. Six synchronous vibration channels are segmented into fixed-length windows, standardized using source-domain statistics, and classified by a compact 1D CNN backbone with and without squeeze-and-excitation (SE) channel attention. A deeper ResNet1D baseline is further introduced to examine whether increasing backbone capacity improves cross-bearing generalization under the same source-only training protocol. We compare random segment-level splits with bearing-level cross-splits that hold out entire bearings as unseen target domains, and we report deployment-oriented indicators including balanced accuracy, false-alarm rate (FAR), and miss rate over five random seeds. Under random splits, the compact CNN baseline reaches near-ceiling test accuracy, confirming that the benchmark is readily separable under in-domain interpolation. In contrast, cross-bearing evaluation reveals severe degradation: in the representative split, the baseline CNN accuracy collapses to approximately 15% with near-zero normal-class recall, while ResNet1D improves fault sensitivity but still retains a high FAR above 88%. Additional cross-bearing permutations further show that this degradation is not attributable to a single unfavorable source–target split. These findings indicate that, under the tested source-only backbones and protocols, distribution mismatch is a dominant bottleneck for deployment-ready cross-bearing diagnosis. The results establish a reproducible baseline for protocol-driven evaluation in aero-engine PHM and motivate future work on domain adaptation, domain generalization, calibration, and sequential decision logic. Full article

Industrial control systems (ICS), which manage critical infrastructure such as power grids and water treatment, are increasingly exposed to cyber threats and operational faults as their connectivity to external networks grows. AI-based anomaly detection has emerged as a key defense, yet three limitations restrict its practical deployment: (i) detected anomalies are treated uniformly without distinguishing between transient faults and intentional attacks, hindering tailored incident response; (ii) the trade-off between detection accuracy and the false-positive rate burdens experts with extensive manual triage and delays prompt action; and (iii) prevailing feature-attribution Explainable AI (XAI) techniques such as SHAP and LIME produce fragmented sensor-level explanations and fail to capture correlations among sensors in time-series data, undermining trust in model decisions. To address these gaps, this paper proposes a graph-based deep learning framework that (a) defines anomaly types in terms of the anomalous-sensor ratio measured before and after smoothing—which operationalizes the correlation-maintenance principle that faults keep coupled sensors jointly anomalous while attacks isolate them—enabling explicit separation of faults, attacks, false positives, and false negatives; (b) identifies ambiguous decisions near the detection threshold as candidate false alarms via dynamic threshold smoothing; and (c) provides correlation-aware graph visualizations for intuitive interpretation. Experiments on the Secure Water Treatment (SWaT) dataset center on this post-detection layer: built on a standard graph-based detector (F1-score 0.787 at Top-K = 10) that serves only as the substrate, the categorization separates faults from attacks, and the subsequent ambiguity analysis identifies false negatives with 83% precision and false positives with 73% precision. By separating attacks from faults and surfacing high-likelihood false alarms together with intuitive sensor-correlation explanations, the proposed approach reduces analyst workload and supports more reliable, prioritized incident response in ICS environments. Full article