The study analyzes pedestrian behavior in multidirectional flow scenarios through controlled experiments conducted in Ghana and China to understand cultural differences. Researchers captured and evaluated both individual and group pedestrian movements in a uniform circular environment across different population densities. Data were collected using PeTrack and analyzed in MATLAB and OriginPro to extract trajectories and calculate key parameters, including average speed, movement time, detour coefficient, and path length. A larger group cohesion and movement synchronization was observed among Chinese participants, while spontaneous and individualistic behavior was observed among Ghanaians. As densities increased, people exhibited more detour behavior and wait-and-follow strategies along with increased path deviations. No statistically significant difference in speed was observed between individuals and groups, but social groups generally traveled longer paths due to coordination constraints. Culturally unique behavioral patterns provide essential information to refine crowd simulation models and create evacuation plans that respect cultural differences. The research outcomes will improve crowd simulation strategies across diverse environments through culturally specific evacuation models.
This paper proposes an advanced settingless protection scheme for AC microgrids based on dynamic state estimation (DSE) and confidence-level analysis. Unlike conventional adaptive protection methods that rely on predefined relay settings, pick-up thresholds, and coordination time delays, the proposed approach eliminates manual settings entirely by continuously evaluating the dynamic health of the protected components. A 7-bus AC microgrid is modeled in a real-time digital simulator (RTDS), and both static state estimation (weighted least squares (WLS)) and dynamic state estimation (extended Kalman filter (EKF)) are implemented using phasor measurement units (PMU) measurements under steady-state, contingency, and fault conditions. Simulation results demonstrate that static estimation fails to track fast transients during faults, whereas EKF accurately captures system dynamics in both grid-connected (GC) modes and islanded (IM) modes. By integrating confidence-level evaluation with DSE outputs, the proposed method enables fast, autonomous, and reliable fault detection without relay coordination, thereby validating its effectiveness for protecting highly dynamic microgrid environments.
The increasing integration of wind power into the electrical grid requires accurate power generation forecasting to maintain the grid’s stability and efficiency. Wind, as a crucial renewable source, poses challenges due to its inherent volatility. Accurate wind speed and direction prediction is both a technical hurdle and an economic imperative for maximizing wind power efficiency and minimizing costs by tailoring production to meet demand. However, current methods often face limitations due to complexity or lack of interpretability. This paper proposes a novel two-stage hybrid transformer-based wind power forecasting model, which combines the strengths of traditional statistical methods and advanced machine learning techniques. The first stage focuses on highly accurate wind parameter estimation, while the second one leverages these estimates for energy production forecasting. This hybrid architecture elevates the temporal capabilities of transformer models enhanced by incorporating domain-specific knowledge and statistical features, and offers a robust solution for enhancing grid management and operational planning. The proposed model is evaluated on multiple wind power datasets, demonstrating significant improvements in forecasting accuracy compared to the existing methods. This hybrid transformers model approach outperforms the existing models by a substantial margin, reducing mean squared error by over 74.8
C–Mn pearlite-type high-strength low-alloy (HSLA) steel proves to be the preferred material for demanding applications such as pressure vessels, heavy-duty vehicles, naval vessels, modern submarines, and offshore platforms. The material demonstrates excellent performance with high strength, exceptional sub-zero impact toughness, acceptable ductility, while maintaining good corrosion resistance and excellent weldability. The automated ball indentation (ABI) technique was employed to assess the mechanical properties of steels across weld joints. This method enables the determination of yield strength, ultimate tensile strength, strain hardening exponent and strength coefficient, providing a comprehensive evaluation of the material’s mechanical behaviour. This study investigates the microstructure and variations in mechanical strength across HSLA steel weld joints produced using both hot wire gas tungsten arc (HWGTA) welding and conventional gas tungsten arc welding at room temperature. The base metal microstructure consists of ferrite and pearlite, while the heat-affected zone (HAZ) and weld metal are characterised by bainite and acicular ferrite, respectively. Notably, the weld zone exhibits a comparatively coarser acicular ferrite microstructure due to the increased heat input in HWGTA welding. ABI testing was conducted at 25°C on the weld region, HAZ region and base metal of both the weld joints. The flow curves derived from the ABI measurements showed strong correlation with those obtained from conventional tensile tests. Key mechanical properties including yield strength, ultimate tensile strength and strain hardening exponent, were determined for each region of the weld joints and systematically compared. The HAZ demonstrated superior tensile strength compared to other zones, which is attributed to the formation of bainite in this region. Overall, the ABI-derived results demonstrated good correlation with those obtained through standard conventional testing methods, confirming the reliability of the ABI technique for localised mechanical characterisation of welded structures.
AISI 52100 bearing steel is widely used in high-load, cyclic loading applications where wear and surface fatigue are critical concerns. This study examines the impact of high power impulse magnetron sputtering (HiPIMS-Cr) pretreatment on coating-substrate adhesion before depositing TiAlN (Ti/TiN interlayers) multilayer films via direct current magnetron sputtering (DCMS), and compares the results with those of conventional Ar pretreatment used in commercial PVD recipes. Coatings were characterized using SEM, EDS, XRD, Rockwell indentation, and scratch testing. XRD analysis revealed that HiPIMS-Cr pretreatment modified the preferred crystallographic orientation from a dominant (111) to competitive growth along (111), (200), and (220) reflections. Ar-pretreated films showed finer grains ( 43 nm) and higher hardness (27.1 GPa), while Cr-pretreated coatings had rougher, cauliflower-like morphology ( 73 nm) and slightly lower hardness (24.3 GPa). Despite this, Cr-treated films demonstrated a 17
The process of flow laminarization at low Reynolds numbers plays an important role in aerospace engineering and in the design of jet and rocket engines. Laminar flow arises when the boundary layer of the flow is stabilized, and hence the transition from laminarity to turbulence is delayed. Extensive research has been carried out over many decades to achieve laminarization of flow. The prospect of modelling such a fluid dynamical system analytically depends on the accessibility of appropriate analytical fluid models for the fluid-dynamic flow acting on the pipe. This problem is a typical physical model in mathematical fluid dynamics studied extensively in the past decades. In this study, we prove that fractal dimensions play a crucial role in the laminarization of an almost incompressible flow with the generalized Chaplygin gas without subjecting the fluid to streamwise body force or imposing any external gradient pressure or force. In the downstream portion of the porous tube, the entire flow undergoes a transition to turbulence then to relaminarization. This has motivating impacts in aerospace flights vehicles engineering.
A multi-proxy signature scheme enables a designated group of proxy signers to generate a valid signature on behalf of an original signer. Such schemes are useful in distributed environments where delegation, accountability, and joint authorization are required. Among the existing constructions, the scheme proposed by Qingshui Xue and Zhenfu Cao is a representative classical (modular-arithmetic-based) multi-proxy signature mechanism. In this paper, we revisit their scheme and present an improved variant based on elliptic curve cryptography. The proposed construction preserves the delegation structure and security properties of the original scheme, while replacing the underlying arithmetic operations with elliptic-curve-based computations. As a result, the scheme achieves reduced key sizes and lower computational cost for an equivalent level of classical security. We formally describe the protocol, prove its correctness, analyze its resistance to known attacks, and provide a comparative evaluation against the original Xue–Cao instantiation.
This study aims to investigate coordinated control operations for large-scale wind energy conversion systems (WECS) employing permanent magnet vernier generators (PMVGs). To do this, an integral backstepping control scheme with an improved adaptive function for pitch angle adjustment and velocity tracking is first designed. Next, the integral error function is presented to ensure the output power stabilization of PMVG-based WECS. The proposed control scheme enables the generator to maintain an optimum speed throughout various wind conditions. In addition, an improved integral adaptive backstepping controller (IABSC) scheme is introduced for pitch angle regulation, effectively limiting generator overspeed and ensuring stable output power in Region III. The IABSC not only simplifies control implementation but also enhances system robustness, replacing traditional controllers. Then, the stability of the closed-loop system is analyzed using the Lyapunov stability theory. Finally, the simulation results for 20 MW PMVG-based WECS operating under Region II and Region III wind speeds are presented to highlight the superior performance and applicability of the proposed methods compared to existing schemes.
The present study proposes a new metal structural fuse called S-Shaped Plate Dampers (SSPDs) to evaluate the cyclic behavior of concentrically braced frames (CBFs). For this purpose, commercial Abaqus software is used to model the finite elements of the steel frames that can be connected to the S-shaped yielding metal damper. First, a sample of the experimental model of the S-shaped yielding metal damper, which has been worked by previous researchers, is verified, and then this damper is installed in a suitable position of the braced frames. After equipping the braced steel frames with a damper, the steel frame system is subjected to cyclic loading and its cyclic performance is evaluated. This damper consists of two S-shaped plates made from ordinary steel and is easy to construct, install, inspect, and replace. A total of ten specimens were subjected to cyclic loading to investigate the failure mode and cyclic performance. The results indicated that with the rise in the material thickness and strength of s-shaped plates dampers, the cyclic response became more stable, the energy dissipation was doubled, the viscous damping grew by 43.48
We investigate a stage-structured predator–prey framework incorporating fear in prey due to adult predators, prey refuge, cooperative hunting among adult predators, and counterattacks by prey against juvenile predators. The model takes into account nonlinear functional responses, handling time, and refuge effects, as well as intraspecific competition in both prey and predator groups. We establish positivity and boundedness, determine the existence of ecologically significant equilibrium points in the system, and investigate their local stability properties using linearisation and eigenvalue analysis. Through bifurcation analysis, we identify that the cooperative hunting coefficient, prey refuge, and predation rate each lead to Hopf bifurcations, while the prey counter-attack rate induces a transcritical bifurcation. Numerical simulations confirm the analytical findings, pointing out the important roles of prey refuge and cooperative hunting in determining population persistence and system stability. In addition, the model allows investigation of the interaction between these two factors, providing insights into their joint influence on the dynamics of the predator–prey system.
A compact Ultra-Wideband Dual Notched Antenna (UDNA) of size 0.368 λ×0.286 λ× 0.0105 λ, is presented and analyzed where, notation lambda is the wavelength corresponding to the lowest frequency. The UDNA covers a wide spectrum spreading across 2.6–8.3 GHz while incorporating two notches at 3.23–4.03 GHz and 5.2–5.8 GHz so as to mitigate the electromagnetic interference and signal attenuation issues originating from wireless networks like WiMAX and WLAN spectrums. The notch bands are achieved by incising a pair of slots engraved onto the radiator in the shape of letter ‘U’. To upgrade and boost the antenna’s radiation without compromising bandwidth, a mono-layer FSS structure featuring a unit cell size of 0.29 λ × 0.29 λ × 0.022 λ is deployed underneath the antenna. This integration results maximum gain of 9.1 dBi with high radiation efficiency. The antenna characteristics are analyzed in free space and in ground coupling sub surface scanning mode with and without FSS by EM wave simulations which have been validated by experimental measurement. The UDNA-FSS exhibits a linear transfer function and consistent group delay across the spectrum, excluding the notch bands, thereby ensuring minimal signal dispersion and reliable subsurface imaging capability. An electro-magnetic sub-surface scanning experiment is also conducted where the result establishes the superiority of performance by integrating FSS with the UDNA.
Maintaining stable and efficient load frequency control (LFC) in interconnected power systems is crucial, particularly in the presence of system nonlinearities and load variations. Conventional proportional-integral (PI) and sliding mode control (SMC) controllers often struggle with fixed-gain constraints and chattering effects, which can compromise overall system performance. This study introduces an advanced sliding mode reaching law (ASMRL)-based SMC controller for LFC in a two-area interconnected power system, incorporating a state-dependent power term and a checkmark function to minimize chattering while maintaining strong control capabilities. Rather than using the conventional sign function, the ASMRL employs the hyperbolic tangent function, thereby enhancing the stability and adaptability of the system. Simulation results confirm that the ASMRL controller achieves quicker convergence, lower peak deviations, improved robustness, and reduced settling time compared to traditional sliding mode controller (TSMRL), integral reinforcement learning (IRL), and PI controllers, particularly under uncertain load disturbances and renewable energy sources. Furthermore, the ASMRL controller efficiently addresses governor dead band (GDB) and generation rate constraints (GRC), making it a viable approach for smart grid frequency regulation.
This study presents the development and validation of an efficient computational tool tailored for analyzing transient reacting flows, with a primary focus on stationary flames. Leveraging the Reynolds-averaged Navier–Stokes equations and employing the k-ε turbulence model, the tool was validated using various single-step chemical kinetics models, specifically simulating methane combustion in a Sandia D flame case. The eddy dissipation concept (EDC) within the OpenFOAM computational fluid dynamics code emerged as the most effective model, surpassing existing eddy dissipation model and infinitely fast chemistry alternatives by 42 K and 121 K, respectively. Further refinement involved applying corrections to the k-ε model, with Pope’s correction exhibiting the highest success, resulting in a 124 K overprediction of temperature and a 0.78 m/s underprediction in velocity compared to experimental results. In-depth investigations into variations within the EDC combustion model revealed the superiority of the EDC 81 model, showcasing accurate predictions for axial velocity, temperature, and product species. The outcomes of this study offer a comprehensive computational framework for analyzing transient reacting flows, with implications for optimizing combustion simulations and advancing understanding in stationary flame dynamics.
Photocatalytic reduction of Ni2+ from wastewater represents a sustainable approach for both heavy metal removal and green hydrogen generation. This study investigates the integration of fluid dynamics, radiation transport, and photochemical reactions in a solar-illuminated batch reactor using a CuFe2O4/TiO2 hetero-junction. Ni2+ ions are reduced to metallic state Ni0-clusters, which simultaneously enhance hydrogen production via water splitting. Simulation results indicate that catalyst dosage, reactor geometry, and light intensity critically influence the Local Volumetric Rate of Energy Absorption (LVREA) as well as Ni2+reduction, and H2 yield. Maximum Ni2+ removal was achieved at low concentrations (10–30 ppm), while H2 production reached 9 ppm after 125 min under optimal conditions. Coupled Computational Fluid Dynamics (CFD) and radiation modeling were performed using ANSYS Fluent with a User-Defined Function (UDF) to describe photocatalytic reaction kinetics and solar energy input. The laminar Navier–Stokes equations, species transport, and Discrete Ordinates (DO) radiation model were solved. Radiation absorption and scattering coefficients were defined on the basis of CuFe2O4/TiO2 properties, and reaction rates were modeled using Arrhenius-type expressions and LVREA. Two- and three-dimensional geometries were meshed with Gambit, and the results were validated against theoretical and experimental Ni2+ reduction profiles. The model also accounted for the Snell’s law, Beer–Lambert attenuation, and optical boundary conditions across glass–water–air interfaces.
Groundwater is the primary source of drinking and irrigation water in Patna, Bihar, where rapid urbanization and intensive agriculture have raised concerns about its quality. Fluoride contamination, in particular, poses severe health risks, including dental and skeletal fluorosis. The aim of this study was to evaluate groundwater quality and predict fluoride concentrations using a combined conventional and machine-learning approach. A total of 100 groundwater samples were collected from urban and suburban areas, and physicochemical parameters (pH, electrical conductivity (EC), total dissolved solids (TDS), major ions, and heavy metals) were analysed. The entropy water quality index (EWQI) was calculated to assess overall water quality, and most samples were classified as ‘excellent’ or ‘good’, with localized instances of contamination. For predictive modeling, three ensemble techniques – decision tree (DT) regressor, random forest (RF), and gradient boosting regressor (GBR) – were applied to estimate EWQI, with RF achieving the highest accuracy (R2 and Nash–Sutcliffe efficiency (NSE) = 0.9706). For fluoride prediction, six models were tested, including optimized RF, DT, XGBoost regressor (XGB), GBR, support vector regressor (SVR), and multiple linear regressor (MLR). Feature importance analysis refined input variables, enhancing prediction accuracy. The random forest (RF-R) emerged as the most reliable model (R2 = 0.950; mean squared error (MSE) = 0.000754; root mean squared error (RMSE) = 0.027459), followed by DT-R and XGB-R. These findings demonstrate that integrating conventional indices with machine learning provides a robust framework for groundwater quality monitoring and targeted fluoride risk assessment. The originality of this study lies in combining EWQI with ensemble learning to achieve high-precision predictions, offering a scalable approach for environmental management in fluoride-prone regions.
With the increasing population, it is obvious that confronting frequently occurring gathering events raises security issues for human beings. Surveillance cameras and security personnel play an important role in keeping track of individuals’ behavior. However, in more crowded scenarios, it is more difficult to monitor individuals’ activities. Therefore, there is a need for intelligent video surveillance systems to analyze crowd behavior. In this review, we have focused on crowd and group behavior analysis in anomalous and non-anomalous situations. Crowd behavior analysis is defined by the appearance and motion pattern analysis of objects. The objects may be singular or a whole crowd, depending on microscopic and macroscopic approaches, respectively. Objects having the same appearance pattern may have abnormal behavior because of different motion patterns, and vice versa. Therefore, we divide the analysis of crowd behavior into two categories: appearance-based and motion-based pattern analysis. Appearance pattern analysis is concerned with the detection and localization of an object’s physical properties, such as shape, dimensions, and location coordinates. Motion pattern analysis, on the other hand, describes methods for analyzing motion parameters such as velocity, acceleration, speed, motion vectors, diversion, curl, and many more. This review aims to present a comprehensive framework for analyzing anomalous crowd behavior using conventional machine learning (ML) (non-deep learning (DL)) techniques, which continue to be efficient and applicable for real-time and resource-limited surveillance contexts. Recent DL-based methods have shown good results, but they usually need big datasets and a lot of computing power. Therefore, this review aims to demonstrate that conventional ML-based technologies still offer ample opportunities for exploring this topic.
Sub-cooled flow boiling in confined geometries plays a pivotal role in thermal management of nuclear reactors and high-flux heat-exchange systems. This study developed and validated a high-fidelity computational framework employing a coupled Eulerian–Eulerian two fluid model (EE-TFM) and a discrete homogeneous population balance model (PBM) to simulate sub-cooled boiling flows under both high- and low-pressure conditions. The framework incorporates mechanistic wall boiling models based on heat flux partitioning, alongside detailed interfacial force modelling including drag, lift, wall lubrication, turbulent dispersion, and virtual mass forces. The PBM accounts for dynamic bubble population evolution via nucleation, coalescence, breakup, binary breakup, and drift mechanisms, enabling accurate prediction of bubble dynamics. Simulations are performed using OpenFOAM and validated against benchmark experimental datasets, demonstrating strong agreement with experimental results. The proposed approach enhances the predictive capability of sub-cooled boiling models and offers a computationally tractable yet physically rigorous tool for thermal hydraulic analysis in nuclear safety and energy systems design.
The telecom industry in India is growing rapidly, and the Bureau of Indian Standards (BIS) has released a code of practice, IS 17740 (2022), for telecom structures. Prior to the release of this standard, engineers followed guidelines mainly from transmission line practices (IS 802 Part 1/Sec. 2) combined with general standards for wind (IS 875 Part 3), referred to in this paper as the prevailing design practices. Numerous structures were built across the nation using the prevailing design practice prior to the release of IS 17740. This paper focuses on the implications of the guidelines prescribed in IS 17740 for the prevailing design practices. Three different telecom tower configurations are considered, representing practical scenarios, to study the impact of IS 17740 on their structural response: a 60-m-high square angular tower, a 60-m-high triangular hybrid tower, and a 30-m ground-based monopole. Based on the case study, an average increase of 6–7
Individual spectral matching techniques, namely Spectral Angle Mapper (SAM), Jeffries-Matusita (JM) along with Spectral Correlation Mapper (SCM) are used to extract enhanced information from hyperspectral image data. SCM has a propensity to overestimate the strength of spectral match in some cases, resulting in inaccurate predictions of biophysical parameters. To reduce such an overestimation, this research provides a hybrid approach, combining two stochastic metrics: Jeffries-Matusita distance measure and Spectral Correlation Mapper. The developed JM-SCM algorithms are implemented to characterize the complex mangrove ecosystems, namely Pichavaram and Muthupet in southern India, using EO-1 Hyperion imagery. The image-derived reference spectra were used for matching Avicennia sp., Rhizopora sp., paddy, groundnut plantation, mudflat, sand, turbid and clear water (in Pichavaram); and Avicennia sp., Prosopis juliflora, mudflat, marshland, saline soil, turbid and clear water (in Muthupet). In the matching-based classification of Pichavaram imagery, the combined JM-SCM (tangent and sine) approaches provided significantly improved accuracies of 93.50
In this study, a 3D thermo-mechanical finite element model was developed in Abaqus/Explicit to simulate the friction stir welding (FSW) of AA7075-T6 aluminium alloy, incorporating a Johnson–Cook constitutive law to represent the material’s viscoplastic behaviour. The model was validated against experimental temperature measurements and microstructural observations. Two tool geometries, a flat-shouldered tool and a scrolled-shoulder threaded tool, were compared under rotational speeds ranging from 1100 to 1700 rpm and welding speeds between 70 and 120 mm/min. The simulations revealed that rotational speed is the dominant parameter influencing heat generation, with the peak temperature rising by approximately 250°C as the rotational speed increased from 1100 to 1700 rpm. Conversely, increasing the welding speed from 70 to 120 mm/min reduced the peak temperature by about 100–130°C, due to shorter tool–workpiece interaction times and reduced frictional heat input. At 1320 rpm and 70 mm/min, the maximum temperature reached 253°C numerically and 187°C experimentally, showing good qualitative agreement between model and experiment. The use of a scrolled-shoulder tool produced higher and more uniformly distributed temperatures, improving material flow and reducing surface defects compared with the flat-shouldered design. Overall, the validated model accurately captured the thermal behaviour of FSW, demonstrating that optimised tool geometry and process parameters can significantly enhance joint quality, refine the stir zone microstructure, and minimise thermal gradients in high-strength aluminium alloys.
