This study establishes a broadband coupling model of "gas-insulated substation (GIS)-grounding grid-soil" using Current Distribution, Electromagnetic Fields, Grounding and Soil Structure Analysis (CDEGS) software to mitigate the risks associated with transient ground potential rise (TGPR) during GIS switching operations, thereby protecting equipment insulation and ensuring personnel safety through simulation calculations. Simulations were predominantly performed considering parameters such as soil resistivity, switch position, metal enclosure height above ground, metal bracket cross-section, non-uniform grounding grid configuration, and vertical grounding electrodes. Subsequently, a quantitative analysis was conducted on the distribution characteristics of transient ground potential rise under different conditions. Research findings suggest that TGPR caused by GIS switching operations is predominantly localized around bushing grounding locations. It increases considerably with rising soil resistivity and greater metal enclosure-to-ground distances, although the rate of increase diminishes in later phases. At the same time, expanding the cross-sectional area of metal brackets effectively diminishes peak values and enhances the distribution of potential. Furthermore, employing appropriately spaced grounding grid layouts serves as an effective supplementary measure to suppress TGPR. These findings offer important guidance for the design of GIS grounding systems in practical engineering applications.
Outsourcing encrypted data to the cloud demands efficient and confidentiality-preserving search techniques, yet existing Searchable Symmetric Encryption (SSE) schemes suffer from severe access and search pattern leakage, and the problem is more intractable in multi-data-owner scenarios where cross-dataset privacy-preserving search remains unsolved. This paper proposes PSSE-MDO (Privacy-Preserving SSE with Pattern Obfuscation for Multiple Data Owners), the first SSE scheme that achieves robust access and search pattern hiding for multi-data-owner settings with constant client-side overhead. Our core innovation is a cryptographic protocol enabling users to issue a single constant-sized search token (trapdoor) for searching across all data owners' encrypted indexes. The scheme integrates Inner-Product Predicate Encryption (IPE) and Somewhat Homomorphic Encryption (SHE) to probabilistically obfuscate search results for each query. A key experimental result shows that the client-side cost approximately 120 ms for search token generation—remains constant regardless of the number of data owners, in contrast to existing pattern-hiding schemes with linearly scaling costs. This validates PSSE-MDO's practicality for real-world large-scale collaborative cloud systems, marking a significant step toward practical privacy-preserving search in federated cloud environments.
paper presents an adaptive & micro;quasilogarithmic gradient quantization framework aimed at reducing memory and computational demands during deep neural network training. The approach employs a companding function with a dynamically adjusted & micro; parameter that adapts to the statistical properties of gradients. Two quantization strategies are developed: a switching & micro;-quantizer that toggles between low-bit uniform and high-bit quasilogarithmic modes based on gradient variance, and a hybrid & micro;-quantizer that statically applies uniform quantization to small gradients and quasilogarithmic to larger ones. Experiments on multiple-layer perceptron (MLP) and convolution neural network (CNN) models trained on CIFAR-10 show that both quantizers retain classification accuracy close to full-precision (FP32) baselines while significantly reducing gradient reconstruction error (RMSE). The hybrid variant consistently achieves better validation accuracy, lower RMSE, and faster convergence than the switching scheme. These results highlight the potential of hybrid quasilogarithmic quantization as an efficient and scalable solution for training deep models in memory or bandwidth constrained environments.
This paper proposes an integrated thermal management framework for battery electric vehicle (BEV) propulsion systems, combining five coordinated strategies into a unified control architecture. The approach includes thermal Model Predictive Control (TMPC), residual heat recovery from the motor and inverter, variable-speed actuation of the pump-fan-chiller, a "battery-first" thermal distribution logic, and HVAC-assisted thermal preconditioning. A coupled electro-thermal RC network model is developed forAa liquid-cooled PMSM traction motor, automotive inverter, NCM battery pack, and cooling circuit, with thermal limits and derating laws aligned to OEM practice. Dynamic simulations on the WLTP Class 3b cycle are performed for four ambient temperatures (-10, 0, 25, 40 degrees C), capturing transient temperatures, temperature-dependent losses, heat flows, and auxiliary energy demand. Results show that all components remain well below their thermal limits, with no derating activation. The analysis demonstrates that range degradation in non-nominal climates is driven primarily by HVAC energy demand and battery charge-acceptance limitations, while propulsion losses remain nearly insensitive to ambient temperature. The proposed framework highlights the central role of intelligent, integrated thermal management in maximizing efficiency, safety, and real-world BEV range.
the era of multi-cloud environments, the challenge for end users or organizations is to select the most appropriate cloud service provider (CSP) due to the complexity and variability of service-level offerings by various CSPs. In this paper, we propose a MAGIQ-TOPSIS-Fuzzy-logic based Cloud Service Provider Ranking (MTF-CSPR) model, which provides aAmore robust and intelligent decision-making system in CSP selection. The proposed model addresses both qualitative and quantitative evaluation criteria by capturing and processing the inherent vagueness of linguistic assessments through fuzzy logic. MAGIQ is employed to derive optimal weights for evaluation criteria that reflect their relative importance towards the alternatives. TOPSIS is applied to rank CSPs where the ideal and anti-ideal solutions are derived, and then based on CSP's proximity to the ideal solution, the ranks are derived. The proposed model ensures accuracy, scalability, and context-awareness for the CSP ranking by integrating subjective expert opinions with objective performance metrics. Experimental evaluation demonstrates that the proposed MTF-CSPR model provides more consistent and accurate rankings compared to conventional standalone approaches, as well as the other ranking models under consideration, providing better decision-support for enterprises to adopt the hybrid or multi-cloud strategies.
heterogeneous wireless sensor networks, hole repair is the core task of ensuring network coverage and connectivity. The key is to accurately locate the best coverage points and efficiently schedule mobile nodes for patching. To this end, this paper proposes an optimal hole coverage point positioning algorithm based on gradient descent method and priority mechanism, and uses the positioning results to guide mobile nodes to perform path planning through deep reinforcement learning algorithm. First, the boundaries of the hole area are positioned, the hole area is reconstructed, the optimal coverage point is positioned using the gradient descent method, and the hole repair priority algorithm is proposed to sort the repair tasks. Secondly, a mobile node scheduling framework based on deep reinforcement learning is built. This framework integrates dual deep Q networks and priority experience replays, so that mobile nodes can not only effectively avoid obstacles during repair, but also achieve trade-offs and coordination between multiple optimization goals such as path length, energy consumption and time. The simulation experiment results show that the use of the hole positioning and repair algorithm proposed in this paper can effectively improve the overall coverage quality of the network and expand the network life cycle.
Partial shading is a major challenge in photovoltaic systems, causing significant power losses, mismatch effects, multiple peaks in the nonlinear power-voltage (P-V) characteristics, hotspot formation, and long-term performance degradation. The severity of these impacts depends on the shading pattern, intensity, and spatial distribution of the affected modules within the photovoltaic (PV) array. To improve energy extraction from PV arrays under non-uniform irradiance, this paper proposes a Simple Physical String Arrangement Reconfiguration (PSAR) technique, which rearranges modules without requiring additional circuitry, sensors, or control algorithms. A 6 & times; 4 PV array was modeled and simulated in MATLAB/Simulink under seven representative partial-shading scenarios. Key performance evaluation indicators, including maximum power, mismatch loss, power loss, fill factor, and performance ratio, were computed and compared with four conventional topologies based on the Series-Parallel configuration: Series-Parallel, Bridge-Linked, Honeycomb, and Total-Cross-Tied. Simulation results demonstrate that PSAR increases energy yield by 10-25% and reduces power imbalance losses by 15-35% relative to the other layouts, whereas it maintains smoother P-V curves. Its simplicity, scalability, and independence from array geometry make PSAR a practical and cost-effective approach to enhance both efficiency and operational reliability in PV systems.
Linear-to-circular polarization converters find applications in modern 5G and 6G communication systems, satellite communications, and the Internet of Things paradigm. A cost-effective three-dimensional frequency-selective surface that uses water as the active material and silicone rubber as the container is proposed. The operation of the design in the 0.1-1 GHz frequency range is assessed through full-wave simulations. Field distribution maps are employed to elucidate the underlying resonance mechanisms. The stability of the proposed structure with respect to the angle of incidence and frequency is evaluated. A quasi-linear variation of the operating frequency with temperature is demonstrated, enabling temperature tunability. The proposed solution is suitable for prototyping, demonstration, and educational purposes in an eco-friendly context.
network traffic is a fundamental task in network management, security, and quality of service provisioning. Traditional approaches usually depend on packet-level content or flow-level statistical features. These approaches require data or packet parsing and maybe constrained in scenarios involving encryption or limited payload visibility. This paper introduces a novel signal-level methodology, where raw Ethernet signals are captured and converted into visual representations for classification using convolutional neural networks (CNNs) with transfer learning. A dataset was constructed from six network protocols by transmitting traffic over a 10Base-T link and capturing the corresponding signals with an oscilloscope. Among the examined representations, scalograms achieved the highest accuracies across CNN architectures, with DarkNet-19 and DarkNet-53 reaching 98.21% and 98.34%, respectively. While deeper models provided slight accuracy gains, they incurred substantially higher training costs. Furthermore, results indicate that greater dataset representativeness improves model performance. Overall, the findings demonstrate that CNNs can effectively learn discriminative features from raw Ethernet signals, enabling high-accuracy traffic classification without packet content and highlighting signal-level methods as a promising alternative to traditional techniques.
High Resolution Range Profiles (HRRPs) playAa critical role in radar based Automatic Target Recognition (ATR) by revealing detailed structural information along the range dimension. In legacy radar systems transmitting unmodulated narrow pulses, low inherent range resolution presents significant challenges for HRRP reconstruction, especially under low signal to noise ratio (SNR) conditions. Spectral Inverse Filtering (SIF) is a recently introduced method that enhances resolution through frequency domain deconvolution, but it remains highly sensitive to noise. This paper proposes a physics-informed Convolution Neural Network (CNN) trained using SIF generated clean HRRPs as supervision. The network is trained on a synthetic dataset generated from randomized target profiles under various SNR levels ranging from-5 dB to 40 dB. Quantitative evaluations using Root Mean Square Error (RMSE), Peak Signal to Noise Ratio (PSNR), and Mean Structural Similarity Index (MSSIM) show that the CNN consistently outperforms SIF, especially under low-to-moderate SNR conditions. Visual comparisons confirm the CNN's ability to suppress noise while preserving key structural features such as peak positions and sidelobes. The results demonstrate that data-driven learning can effectively complement physics-based methods, offering robust, high-fidelity HRRP reconstruction without need to modify the radar hardware or transmitted waveform.
This study investigates the surface-mounted permanent magnet synchronous motors (SPMSMs) according to slot numbers with 20 pole rotor. Based on the optimization process with surface response method (SRM) and finite element method (FEM), each machine is optimized for lower cogging torque and torque ripple, and their electromagnetic field characteristics are comparatively investigated. Besides, by applying the derived circuit parameters to the voltage, torque and mechanical equation, the dynamic characteristic analysis is also performed under position control. While the position is well controlled, the speed, current and torque characteristics are addressed based on the position reference. The studied results show that the 20 pole and 21 slot combination has superior performance in comparison with other machines with 24 slots and 27 slots.
The modernization of power distribution grids through the adoption of microservice architecture significantly enhances the availability, reliability, and scalability required by diverse consumers, including households and industries. This paper presents a transition from traditional monolithic systems to a cloud-based microservice framework, which not only facilitates streamlined deployment on cloud platforms but also leverages precise mathematical models to optimize node distribution and microservice configurations. The proposed architecture enables modular deployment, independent scaling of power applications, and improved fault isolation. Simulation results demonstrate substantial improvements in scalability and performance, showing that the microservice-based approach efficiently adapts to increasing network complexity and concurrent client demands for real-time smart grid operations. The framework also opens new market opportunities by enabling flexible, efficient, and scalable power application deployment in cloud environments.
Internet of Things (IoT) devices face exponential growth daily across diverse sectors. The need for robust security measures to safeguard sensitive data and ensure privacy is imminent. Virtual Private Network (VPN) services emerge as a promising solution, offering encryption protocols to fortify communication channels and secure remote access IoT devices. This research conducts a review and performance evaluation of four prominent VPN services: OpenVPN, WireGuard, being the most common on the market, and Tailscale and ZeroTier, because they can function without access to a public IP address. The main goal is to fortify IoT environments. Exploring their performance parameters, encompassing metrics such as the transfer speed test and latency. The empirical findings from this experiment provide valuable insights into the intricate balance between security, performance, and operational feasibility of these VPN solutions in real-world IoT deployments. This nuanced understanding equips IoT security stakeholders and network administrators with the knowledge to make well-informed decisions about IoT security architecture and infrastructure optimization.
Given the critical role of epoxy resin's thermomechanical performance in the long-term safety of gas-insulated switchgear (GIS), this study focuses on enhancing epoxy resin properties through molecular dynamics modeling in Materials Studio. By considering real components, using bisphenol A-type epoxy resin (DGEBA) as the matrix with 3,3'- diaminodiphenylsulfone (33DDS) or methyl-tetrahydrophthalic anhydride (MeTHPA) curing agents, Al2O3 filler, and phenol (Ph) accelerator, several molecular dynamics models with different components were established. Crosslinking analyses in 0% to 93% were performed via Perl scripts, with thermomechanical properties analyzed across different components and crosslinking degrees in the 250-600 K temperature range. The results indicate that Al2O3 fillers effectively raise the glass transition temperature of epoxy resins while significantly enhancing their thermal conductivity. Higher crosslinking degrees enhance overall mechanical properties, while formulas with fillers and anhydride curing agents show improved specific mechanical indices. Additionally, increased crosslinking degrees elevate molecular chain segment mobility, and filler addition reduces mean square displacement curve slopes. This work provides a theoretical base for optimizing epoxy insulation reliability in GIS applications.
Detecting abnormal event in video is essential for maintaining safety in modern communities. However, due to factors of complex background, large changes in scale, and the randomness of abnormal events, causing abnormal event detection poses significant challenges. To address the issue, we propose an effective sparse pooling adversarial learning framework (SPLF) for anomaly event detection, which integrates self-attention and pyramid features into a unified architecture. Specifically, the network takes video frames as input and employs an efficient U-Net to predict unknown frames. Meanwhile, self-attention mechanism and pyramid pooling features are combined to focus on salient areas and capture moving objects with varying scales. In addition, to evaluate the scores of abnormal events, a multi-scale error pyramid is introduced to improve the accuracy and robustness of the proposed SPLF. The comparison test is conducted on three publicly datasets: Ped2, Avenue, ShanghaiTech and a community scenario dataset. The frame-level AUC (area under curve) achieves 97.5%, 89.2%, 75.1% and 70.2% respectively, reaching a high level. Ablation tests further validate the effectiveness of self-attention mechanism and multi-scale pyramid pooling features. The test results demonstrate that the proposed method can effectively learn action patterns and accurately detect abnormal events in community scenarios.
Mobile networks have been expanding over the years with high-capacity and highly scalable optical transport systems. These systems are designed to help network operators handle the increasing data traffic demands and support the deployment of advanced technologies like 5G, IoT and provide services, such as online learning, social media platforms and telemedicine. Radio-over-fiber (RoF) and microwave photonics technology enable efficient transmission and delivery of high-capacity data traffic in fronthaul networks with minimal signal loss and dispersion. This study presents a reliable Fiber-Free Space Optics (FSO) hybrid architecture incorporating a tunable optical delay line filter (DLF) for dynamic dispersion compensation. The proposed fiber-FSO architecture enhances the signal quality and compensates dispersion up to +/- 106 ps/nm in the optical link. This architecture has been analyzed for various climatic conditions such as clear weather (CA), light fog (LF), moderate rain (MR), and heavy rain (HR). The results show that using the DLF enhances receiver sensitivity by about 1.5 dB in all weather conditions, making the system more reliable for high-speed optical communication.
With the introduction of the 5G technology, transmission of multimedia data is very high which demands data security. Image encryption techniques secure the transmission of image data from the intruder. We proposed a one-dimensional chaotic map whose chaotic behaviour over a wide range of control parameters is verified experimentally. Image encryption algorithm using two level permutations and one level diffusion is proposed in this paper. First level permutation is done on block level using one chaotic sequence. In the second level, entire image is permuted. Row and column diffusion is done using two chaotic sequences. From the experimental results and analysis, it is evident that the proposed image encryption algorithm resists differential and statistical attacks.
Considering the operating limits proposed by ISO 10816-21-Evaluation of Machine Vibration by Measurements on Non-Rotating Parts-for horizontal-axis wind turbine gearboxes, we analyzed the behavior of nearly one hundred gearboxes from three nearby onshore wind farms (similar to 10 sq km) in northeastern Brazil. Each wind turbine is equipped with an identical mechanical vibration monitoring system, comprising ten sensors and nine features per sensor. First, we assessed whether the equipment operated within the ISO-defined limits. Next, we confirmed the trends detected by the ten gearbox sensors exhibited a strong correlation with one another. However, trends among similar pieces of equipment operating under the same conditions were not strongly correlated. An unsupervised correlation analysis using the Fast Fourier Transform (FFT) was conducted for all wind turbines, considering the zone boundary values proposed by ISO. The unsupervised correlation analysis enhances knowledge for more targeted monitoring, achieving an accuracy score exceeding 70%. This approach contributes to the development of a more effective predictive maintenance program.
Cloud computing has become an integral part for many organizations as well as cloud users. There are several cloud service providers hosting numerous cloud services. The challenge for the cloud users is to find the suitable cloud service among all cloud service providers which is most trusted and meets user expectations. The authentic and accurate trust values for such cloud services can help organizations and cloud users to choose appropriate cloud services. In this paper, we propose a Digital Twin and User-Preference Based Trust Computing (DTUPTC) model which is based on evidence-based strategy of trust computation. The novelty of our DTUPTC model is that we get the user inputs on their preferred quality of service (QoS) attributes which will be used for trust score computation and also provide their desired values for those QoS attributes. The weights assigned to each QoS attributes while computing trust score are calculated and assigned automatically without any manual intervention. The comparison results of our DTUPTC model with other models proves that the trust scores computed by DTUPTC model are more accurate than the other models.
