Search Results (1,071)

Shearer trailing cable faults are difficult to distinguish from terminal measurements because their signatures coexist with changes in load, source imbalance, cable temperature, and sensor error. This study presents a physics-guided categorical Mamdani framework supported by a distributed-parameter source, cable, and load model. A total of 180 simulations cover normal operation, phase-to-ground faults, A–B inter-phase short circuits, conductor-resistance degradation, and insulation-path deterioration over multiple severities, locations, motor loads, source imbalance levels, and temperatures. Paired pre-fault and post-fault changes in load-terminal voltage unbalance, source-terminal zero-sequence current ratio, and source-to-load voltage attenuation are mapped to fault path-based membership functions and diagnostic rules. Twenty repeated 70/30 holdouts grouped by base operating condition compare the proposed method with a deterministic threshold tree, radial-basis-function support vector machine, and k-nearest-neighbor classifier. Across these repeated grouped holdouts, the proposed method achieves 99.80% mean accuracy with a standard deviation of 0.63 percentage points on noise-free test cases. Accuracy remains 81.12% and 74.71% under 1% and 2% RMS waveform noise, respectively. These results demonstrate the effectiveness of the proposed framework for interpretable diagnosis of shearer power-core cable faults across the investigated simulated operating conditions and waveform-noise levels. Full article

The controllable line-commutated converter (CLCC), which achieves forced commutation and effectively prevents commutation failures, has seen initial engineering applications in high-voltage direct current (HVDC) transmission systems. This paper analyzes fault characteristics and develops a valve control strategy for single-phase-to-ground faults on the valve side of the CLCC converter transformer. First, mathematical models of the DC-side circuit and its control system are established. Next, based on the fault location and changes in current paths during the fault, the transient characteristics of the valve current are analyzed, accounting for the effects of the control system. Furthermore, a relationship describing how faults affect the extrema of the valve current during commutation is derived. Finally, a method for optimizing the valve control strategy is proposed based on predicting the times at which the valve current reaches its extrema. This prediction is obtained using a current transfer time model derived from an equivalent circuit describing current transfer between the main and auxiliary branches of the CLCC. Simulation results show that the proposed method can effectively reduce the current stress imposed on the arrester by fault-induced abnormal commutation, improve the current waveform during commutation, and enhance the operational stability of the system during faults. Full article

The Litang fault zone, with intense late Quaternary activity, frequent strong earthquakes, and dense landslides along its trend, directly threatens the operational safety of the National Highway 318 (G318) Sichuan–Xizang transportation corridor. Predicting potential earthquake-triggered landslide (EQTL) susceptibility is limited by the spatial heterogeneity of near-fault ground motion coupled with geological/topographic conditions. This study proposes a physics-driven, data-integrated framework that couples curvilinear grid finite-difference (CG-FDM) PGA simulation with a fault-geometry-incorporated random forest model—trained on landslides from the 2008 Mw 7.9 Wenchuan earthquake and adapted to the Litang fault zone via the 71° dip sub-model—for EQTL susceptibility prediction. Validation against historical high-susceptibility landslide zones and unstable slopes confirms the results. Maximum PGA reached 0.611 g with banded distributions along the fault, reflecting significant near-fault, basin, and topographic amplification. High and very high susceptibility zones cover 7.64% of the area, mainly within 5 km of the fault and at intersections, overlapping 57% of historical landslide susceptibility zones and containing 75.0% (by number) and 53.0% (by area) of unstable slopes. The minimum angle between ground-motion vector and aspect suggests enhanced triggering along slope direction. In the G318 Litang section, 13.48% (14.21 km) lies within high–very high susceptibility zones, indicating localized hazards. This framework supports EQTL susceptibility mapping and mitigation for major projects. Full article

This study numerically examines the transverse seismic response of a high-pier, long-span stiff-skeleton arch bridge equipped with transverse steel dampers (TSDs). A three-dimensional SAP2000 model represents the primary bridge members predominantly with elastic beam elements. Nonlinear behavior is concentrated in the spherical steel damping bearings (SSDBs), gap links, and idealized TSD links. A sequential staged discrete screening evaluates TSD yield strength, initial gap, and post-yield stiffness ratio using displacement, elastic force-demand, and hysteretic-response indices. Fifteen recorded motions are then used to examine response trends across pulse-period groups. Nine pulse-like records are also decomposed to compare the original motions with their residual components. Within the predefined candidates, 3000 kN, 150 mm, and 15% provide the selected balance among the adopted response indices. This combination is conditional on the candidate set, stage order, bridge model, and input motions, and is not a global multi-parameter optimum. An additional 15-record check of the 13 unique parameter configurations found that no candidate minimized representative bearing, pier, and arch responses simultaneously. This supports interpreting the retained values as a conditional compromise. Exploratory record-level tests identify false-discovery-rate-controlled group differences for pier-top displacement, arch-rib displacement, and arch-rib bending moment; most force-demand indices do not show resolved group differences in this small set. In the nine paired original–residual comparisons, median peak-response reductions range from 16.8% for pier-base shear to 76.9% for arch-rib bending moment, although the pier-base shear change is not statistically resolved. These findings provide comparative numerical demand trends for the investigated bridge. They do not constitute member-capacity verification, physical-device validation, or a general seismic-safety assessment. Full article

To improve voltage support at the point of common coupling of renewable energy collection stations connected to weak grids, this paper proposes a synchronous machine-like outer reference generator for a conventional fixed-PQ voltage-source converter (VSC). The proposed versatile static synchronous machine (VSSM) emulates the excitation and reactive power-voltage regulation of a synchronous generator, maps the resulting transient internal electromotive force to an exact reactive power reference, and applies converter capability and ramp rate limits with an equivalent Q-tracking response. A two-stage strategy further uses fast dynamic compensation to clamp voltage and slower static compensation to release bidirectional dynamic reserve. In the WSCC nine-bus benchmark, all controllers use the same 100 MVA VSC, +/−80 Mvar limit, 0.04 s reactive power inner loop, and 10 pu/s ramp limit. After a three-phase fault and line tripping, the minimum post-fault bus 8 voltage is 0.9452 pu with the VSSM, compared with 0.8755 pu for the calibrated PI-SVG and 0.7660 pu for fixed Q; Q-U droop gives 0.9492 pu but requires 0.33 s for sustained recovery above 0.95 pu, compared with 0.23 s for the VSSM. A positive-, negative-, and zero-sequence extension also evaluates a single-line-to-ground fault. Finally, an anonymized actual receiving-end grid model shows an approximately 40 ms improvement in voltage recovery and reduced steady-state loading of dynamic reactive sources. All results are obtained from offline simulations. Full article

To address the problems of existing DC crosstalk fault detection methods, including complex procedures, difficult implementation, and potential insulation hazards, this paper proposes the crosstalk fault detection and location method based on variable bridge. First, by periodically switching the busbar-to-ground resistance through a variable-bridge circuit in one DC system, a low-frequency AC signal is injected into one busbar, while the bus voltage of another DC system is synchronously detected. If this low-frequency AC signal is present, it can be determined that a crosstalk fault exists between the two DC systems; this signal injection method avoids the insulation safety hazards caused by the coupling capacitor injection method. Subsequently, a method of alternating injection and detection between the two system sections is adopted to locate the crosstalk branch. Finally, the dual-frequency method is introduced; namely, AC signals of two different frequencies are injected and the impedance of the crosstalk loop is measured separately, and a system of two equations is established to solve for the crosstalk resistance. The simulation experiment results show that, whether positive–positive pole crosstalk or positive–negative pole crosstalk occurs, this method can accurately identify the crosstalk branch and measure the crosstalk resistance. Full article

Intelligent bearing diagnosis and maintenance require question-answering systems that combine asset-specific condition evidence with engineering knowledge. However, existing bearing diagnosis studies mainly predict fault labels or health states, while document-based RAG systems generally lack access to the condition of a specific asset. Digital twin-based systems provide structured condition data but offer limited support for manual-grounded diagnostic interpretation. Consequently, there is still no unified route-aware framework and benchmark for determining the evidence that is required by different diagnostic questions and evaluating route selection, DT querying, document retrieval, and answer grounding separately. To address this gap, this paper proposes a route-aware digital twin (DT) and retrieval-augmented generation (RAG) large language model (LLM) agent framework for bearing diagnosis question answering (QA). The framework routes questions to four paths: unrelated, DT-only, RAG-only, and DT+RAG. A route-specific 140-question benchmark is constructed from run-to-failure bearing data and a technical maintenance manual corpus to evaluate the router and the four processing paths. The hybrid router achieves an accuracy of 0.979, while the DT-only module obtains 0.967 query accuracy and 0.900 joint accuracy. Hybrid + reranker performs best for RAG-only Hit@5, mean reciprocal rank (MRR), and context precision, whereas hybrid performs best for DT+RAG Hit@1, MRR, faithfulness, and context precision. These results demonstrate that route-aware evidence selection provides a transparent and evaluable approach to LLM-based bearing diagnosis QA. Full article

Accurate and reliable spud height estimation is fundamental to autonomous cutter suction dredging, where conventional sensors may be affected by harsh marine conditions, degradation, and single-point failures. This paper presents a fault-tolerant vision–ToF sensing framework combining a stereo-camera subsystem, discrete Time-of-Flight (ToF) reference-event detection, linear Kalman filtering, adaptive measurement weighting, and online two-point calibration. A laboratory platform implemented on an NVIDIA Jetson Orin Nano (Santa Clara, CA, USA) was evaluated using eight programmed software-injected measurement faults, with repeat trials for camera_noise and camera_dropout. For the three reference-event-path fault scenarios, peak estimator-reported position variance remained below the baseline peak of 0.87 px², whereas camera-side degradation caused substantially larger covariance increases, reaching 85.21 px² in a camera_dropout repeat. Across four camera_noise repeats, adaptive weighting reduced the mean absolute frame-to-frame output change by 17.3% relative to fixed-covariance replay. The complete online pipeline achieved a mean frame-processing interval of 80.9 ms (approximately 12.4 frames/s) across six repeat experiments. Because no continuous independent dynamic ground truth was available, the results characterise estimator-reported uncertainty and tracking behaviour rather than absolute physical position accuracy. The findings indicate that vision is the primary continuous position channel, while ToF provides discrete TOP/MID physical reference events supporting calibration and event-triggered correction. Full article

The intensification of ground pressure and instability of surrounding rock in deep mining are the core safety challenges of metal mines. This article takes the first mining area of the Sishanling Iron Mine from 960 m to 1020 m as the object, and based on the data obtained from the multi-channel microseismic monitoring system for the whole year of 2025, combined with monthly mining and filling parameters, conducts a rock mass risk assessment that couples microseismic activity with the mining and filling process. The results show that microseismic activity and blasting operations exhibit a significant “resonance of the same frequency” response. The event-intensive areas are distributed along the fault zone and the edge of the goaf, and migrate in a directional manner from shallow to deep with the advancement of mining. By comparing the photos of the tunnel damage on site on a monthly basis, it was found that the incident gathering area was highly consistent with the locations of roof collapse and debris support, which verified the accuracy of microseismic positioning. Based on this, a risk discrimination index based on event frequency, energy release rate, and spatial concentration was established to dynamically evaluate the −960 m and −1020 m sections on a monthly basis. The local risk intensity in the −1020 m section was higher, and the risk increased compensatorily when the filling was delayed 2–4 weeks after mining, revealing the key control role of the mining filling coordination rhythm on the stability of the surrounding rock. The coupled evaluation system of mining filling microseismic risk constructed in this study can provide technical reference for active early warning and differentiated prevention and control of ground pressure in deep mines. Full article

Low-voltage networks are designed as solidly grounded neutral networks (TN, TT) or isolated neutral networks (IT). The latter type is used when continuity of supply and effective protection against electric shock are required despite a single ground fault. A characteristic feature of the IT network is the application of insulation monitoring systems, currently officially named Insulation Monitoring Device (IMD). The aim of this device is to signal the first ground fault, and the network can still be powered. This article shows a comprehensive overview of IMD solutions, from historical to contemporary. IMD structures and characteristic features, as well as critical evaluation, are presented, highlighting their advantages and disadvantages. The desired directions for the development of IMDs are indicated to ensure their proper functioning in modern power networks. Full article

Inland waters such as lakes, reservoirs, rivers, fish ponds, and confined hydraulic structures impose turbidity, shallow-water acoustic multipath, dense boundaries, dynamic biological interference, and limited communication on underwater robots. These coupled constraints can simultaneously degrade sensing, localization, mapping, planning, and control. This structured narrative review synthesizes peer-reviewed English-language evidence on underwater navigation, with emphasis on multisensor fusion localization and the transfer of localization uncertainty into risk-aware planning. Visual, acoustic, inertial, velocity, depth, and external positioning measurements are compared by complementarity, observability, degradation modes, and integration cost. Filtering, sliding-window optimization, factor graphs, estimator switching, and hybrid model- and data-driven approaches are evaluated according to accuracy, real-time performance, robustness, and localization credibility. The review examines how covariance, sensing quality, collision probability, energy, platform dynamics, and task requirements affect maps, path costs, safety margins, trajectory tracking, and degraded operation. Evidence from simulation, public datasets, hardware-in-the-loop tests, tanks, and real waters is synthesized conditionally because study platforms, environments, ground truth, and metrics are heterogeneous. Reliable inland-water autonomy requires diagnosable heterogeneous sensing, integrity-aware localization, and coordinated feedback among localization, planning, and control. Mission-level evaluation should consider data validity, fault recovery, and safe completion rather than average localization error or shortest path alone. Full article

DC distribution networks employing a low-current grounding method offer high power supply reliability. However, when a single-pole grounding fault occurs on a feeder, the fault current characteristics are not distinct, making accurate fault identification and feeder protection challenging. Prolonged operation with the fault may cause insulation damage, potentially leading to pole–pole short-circuit faults and escalating the incident. Therefore, equipping the system with reliable and rapid feeder protection is crucial. This paper analyzes the fault current in a two-level VSC-based DC distribution system under single-pole grounding faults, clarifies the relationship between the transient current of the feeder and the line capacitive current, and further investigates the variation patterns of fault transient current and Teager-based signal-feature characteristics. Based on this, a single-pole grounding fault protection method utilizing the Teager instantaneous-current feature is proposed. The Teager energy operator is employed to extract the instantaneous current feature of each feeder, and a protection criterion is constructed using the dimensionless feature ratio between the faulty pole and the non-fault pole to identify the faulty feeder. A simulation model of the DC distribution network is built using MATLAB/Simulink. Simulation results show that the proposed protection method can adapt to various scenarios, including different line faults, fault locations, fault resistances, and noise, while demonstrating satisfactory performance. Full article

In this paper, the solar irradiance is investigated in a part of central Serbia, where a low-cost IoT sensor network was deployed at three locations: a mountain slope, an open field near a village, and an obstructed position in the city center. All three locations are in the same NASA POWER grid cell. After multi-stage quality control, 26,145 valid daytime records were compared to the satellite reference. The satellite assigns identical values to all three positions but the measured mean daytime irradiances are 267.7, 351.5 and 19.6 W/m², respectively. The rural station is in the best agreement with the reference (R² = 0.567); the mountain station suffers from a persistent shading bias (MBE = −138.9 W/m²); the signal at the urban station is attenuated by surrounding buildings and vegetation by a factor of ~12. Also, 25 regression models (gradient boosting, recurrent, convolutional, fully connected and graph-based) were trained on the 25-year monthly NASA POWER record for the same cell. XGBoost obtained R² = 0.998, the hybrid TCN-GNN R² = 0.968 and a plain ReLU network R² = 0.912, and a seasonal model with calendar features was used to reconstruct the satellite reference for two months of 2026 not yet available in the archive. The deployed hardware, network and energy behavior are documented quantitatively: per-site link delivery ratios of 96.8%, 83.5% and 58.2%, the photovoltaic harvesting record of the nodes, and the absence of any energy-aware transmission scheduling. Two of the trained models were compiled for the ESP32 nodes and measured on the deployment hardware: a depth-limited gradient-boosted corrector runs in 46.3 µs using 11.1 kB of flash, and an INT8 fully connected network in 138.1 µs using 3.0 kB, while the graph hybrid cannot be converted for microcontroller execution at all. This defines an edge–cloud partition in which local inference performs bias correction and fault detection while the cloud tier retains the heavy models and periodic retraining. To the best of the authors’ knowledge, this is the first multi-site ground-based IoT irradiance record for central Serbia with such different terrain types. Full article

Carbonates host valuable groundwater and geo-energy resources in shallow (<0.2 km below groudsurface) and deep (0.2 to 4 km) aquifers, respectively. The hydraulic conductivities of fractured carbonates are unknown at elevated depths (>0.2 km below ground level). To tackle this scenario, the hydraulic conductivity of four Mesozoic units were reviewed in the Umbria–Marche Succession (Italy) due to the large amount of literature available from geothermal exploration and earthquake forecasting. The hydraulic conductivity decreases of three-order (10³) of magnitude from 0.1 to 4 km below the groudsurface according to pumping, and tracer tests, and regional flow models. The Calcare Massiccio and Scaglia Bianca formations appear the most and less hydraulically conductive units, respectively. Seismogenic faults were classified as combined conduit–barriers due to the presence of damage zones and cores. Cores were tested in the laboratory, providing low (up to 8.0 × 10⁻⁸ m day⁻¹) hydraulic conductivities. The small-scale faults studied at shallow depths (0.2 km) can be classified as karstified conduits according to the dataset. Overall, the review showed that a summary of the hydraulic conductivities of fractured carbonates at a variety of depths find application on the sustainable management of shallow groundwater resources and multiple sectors of the geo-energy industry. Full article

Physical sensor deployments in ground-based environmental monitoring networks are frequently constrained by high installation costs, hardware failures, and limited spatial coverage, resulting in incomplete observational datasets and degraded sensing capacity across monitoring stations. Data-driven virtual sensing offers a cost-effective alternative by estimating target environmental parameters through machine learning models trained on correlated sensor measurements, reducing dependency on dense physical infrastructure. This paper presents VS-DCFF, an applied virtual sensing framework for dual-target estimation of near-surface air temperature and relative humidity from ground-based sensor network data. VS-DCFF integrates: (i) a deterministic pipeline applying temporal encoding, rolling-window statistics, and mutual information-based feature selection to capture a linear trend/seasonal component and to select features predictive of the residual signal; (ii) a probabilistic pipeline employing a Gaussian copula model to generate statistically consistent synthetic residual samples preserving inter-variable dependencies; and (iii) an early feature-level fusion strategy feeding a copula-augmented XGBoost residual-boosting stage, whose output is combined with the linear trend component for the final prediction. Under a strict chronological evaluation protocol, VS-DCFF is benchmarked against persistence, linear and ensemble regression baselines, and a same-protocol re-implementation of the statistical core of the VSG-SGL framework, and achieves near-surface air temperature $\mathrm{RMSE}=0.7791$ °C, $R^2=0.9735$, and relative humidity $\mathrm{RMSE}=3.7747\%$, $R^2=0.9701$, outperforming all tested baselines. The framework is further validated through leave-one-station-out spatial generalization, robustness evaluation under simulated sensor faults and target-history loss, copula-variant and synthetic-data fidelity diagnostics, and a lightweight edge-deployment ablation. All findings, including cases where tested extensions such as spatial context features did not yield a robust improvement, are reported transparently. Results indicate that the proposed architecture provides a computationally efficient, extensively validated approach to dual-target environmental virtual sensing under realistic deployment conditions. Full article