Search Results (4,679)

Traditional deep neural networks (DNNs) often suffer from a time-consuming training process, which is restricted by accumulation of excessive network layers and a large amount of parameters. More neural units are required to be stacked to achieve desirable performance. Specifically, when dealing with large-scale datasets, a single DNN can hardly obtain the best performance on the available limited computing resources. To address the issues above, in this paper, a novel Parallel Hierarchical Ensemble Deep Neural Network (PH-E-DNN) is proposed to improve accuracy and efficiency of the deep network. Firstly, the fuzzy C-means algorithm (FCM) is adopted so that the large-scale dataset is separated into several small data partitions. As a benefit of the fuzzy partitioning of the FCM, several sub-models can be obtained through learning their respective data partitions and isolating them from the others. Secondly, the prediction results of each sub-model in the current level are used as the discriminative knowledge appended to original regional subsets, and predictions from each level symmetrically augment inputs for the next level. In the PH-E-DNN architecture, predictions from each level symmetrically augment inputs for the next level, creating a symmetrical flow of discriminative knowledge across the hierarchical structure. Finally, multiple regional subsets are merged to form a global augmented dataset, while multi-level parallel sub-models are stacked to organize a large-scale deep ensemble network. More importantly, only the multiple DNNs in the last level are ensembled to generate the decision result of the proposed PH-E-DNN. Extensive experiments demonstrate that the PH-E-DNN is superior to some traditional and deep learning models, only requiring a few parameters to be set, which demonstrates its efficiency and flexibility. Full article

High-flow aeroengine transient tests involve strong coupling and external disturbances, which pose significant challenges for intake environment simulation systems (IESSs). This study proposes a compound control scheme that combines fixed-time active disturbance rejection with static decoupling methods. The scheme integrates a fixed-time sliding-mode controller (FT-SMC) and a super-twisting fixed-time extended-state observer (ST-FT-ESO). A decoupling transformation separates pressure and temperature dynamics into two independent loops. The observer estimates system states and total disturbances, including residual coupling, while the controller ensures fixed-time convergence. The method is deployed on a real-time programmable logic controller (PLC) and validated through hardware-in-the-loop (HIL) simulations under representative high-flow scenarios. Compared to conventional linear active disturbance rejection decoupling control (LADRDC), the proposed scheme reduces the absolute integral error (AIE) in pressure and temperature tracking by 71.9% and 77.9%, respectively, and reduces the mean-squared error (MSE) by 46.0% and 41.3%. The settling time improves from over 5 s to under 2 s. These results demonstrate improved tracking accuracy, faster convergence, and enhanced robustness against disturbances. Full article

The article discusses a regenerative heat exchange system for a solid oxide electrolyzer cell (SOEC) used in the production of green hydrogen. The heating system comprises four heat exchangers, one condenser heat exchanger, and a mixer evaporator. A pump and two throttle valves have been added to separate the hydrogen at an elevated steam condensation temperature. Assuming steady flow, a thermodynamic analysis was performed to validate the design and to predict the main parameters of the heating system. Numerical optimization was then used to determine the optimal temperature distribution, ensuring the lowest possible additional external energy requirement for the regenerative system. The proportions of energy gained through heat exchange were determined, and their distribution analyzed. The calculated thermal efficiency of the regenerative system is 75%, while its exergy efficiency is 73%. These results can be applied to optimize the design of heat exchangers for hydrogen production systems using SOECs. Full article

In this study, data assimilation using PIV measurement data of the cylinder wake obtained from wind tunnel tests was applied to tune the simulation model parameters of Detached Eddy Simulation (DES) to improve the accuracy of large-scale separated flow simulations around a cylinder. The use of DES enables more accurate simulation of large-scale separation flows than RANS. However, it increases computational costs and makes parameter tuning using data assimilation difficult. To reduce the computational time required for data assimilation, the conventional data assimilation method was modified. The background values used for data assimilation were constructed by extracting only velocity data from locations corresponding to observation points. This approach reduced the computational time for background error covariance and Kalman gain, thereby significantly reducing the execution time of the filtering step in data assimilation. As a result of tuning, $C_{des}$ significantly increased, while $C_{b1}$ decreased. This adjustment extended the length of the recirculation bubble, bringing the time-averaged velocity distribution closer to the PIV measurement data. However, the peak frequency in the PSD obtained from the FFT analysis of velocity fluctuations in the wake shifted slightly toward lower frequencies, slightly increasing the discrepancy with the measurement data. Verifying the relationship between parameter values and flow, it was found that parameter tuning stabilized the separation shear layer generated at the leading edge of the cylinder and enlarged the size of the recirculation bubbles. On the other hand, frequency variations did not show consistent changes in response to parameter value changes, indicating that the effect of parameter tuning was limited under the simulation conditions of this study. To bring the frequency fluctuations closer to experimental results, it is suggested that other methods, such as higher-order spatial and temporal accuracy, should be combined. Full article

To alleviate the pressure on grid regulation and ensure grid safety, pumped storage power stations need to frequently start and stop and change operating conditions, leading to the pump-turbine easily entering the hump characteristic zone, causing flow oscillation within the unit and significant changes in its input power, resulting in increased vibration and grid connection failure. The spatial distribution of energy losses and the hydrodynamic flow features within the hump zone of a pump-turbine under pumped storage operation are the focus of the study. The SST k-ω turbulence model is applied in CFD simulations of the pump-turbine within this work, focusing on the unstable operating range of the positive slope, with model testing providing experimental support. The model test method combines numerical simulation with experimental verification. The LEPR method is used to quantitatively investigate the unstable phenomenon in the hump zone, and the distribution law of energy loss is discussed. The results show that, at operating points in the hump zone, up to 72–86% of the energy dissipation is attributed to the runner, the guide vane passage, and the double vane row assembly within the guide vane system. The flow separation in the runner’s bladeless area evolves into a vortex group, leading to an increase in runner energy loss. With decreasing flow rate, the impact and separation of the water flow intensify the energy dissipation. The high-speed gradient change and dynamic–static interference in the bladeless area cause high energy loss in the double vane row area, and energy loss mainly occurs near the bottom ring. In the hump operation zone, the interaction between adverse flows such as vortices and recirculation and the passage walls directly drive the sharp rise in energy dissipation. Full article

Addressing the technological bottlenecks in the efficient utilization of clay-type Li deposits in China, this study systematically investigates Li occurrence states and develops clean extraction processes using the Jinyinshan clay-type Li deposit in southern Hubei as a case study. The research aims to provide technical guidance for subsequent geological exploration and development of such deposits. Analytical techniques, including AMICS, EPMA, and LA-ICP-MS, reveal that Li primarily occurs in structurally bound forms within cookeite (82.55% of total Li), illite (6.65%), and rectorite (5.20%), with mineral particle sizes concentrated in fine-grained fractions (<45 μm). Leveraging process mineralogical insights, two industrially adaptable blank roasting–acid leaching processes were innovatively developed. Process I employs a full flow of blank roasting–hydrochloric acid leaching–Li-Al separation–Ca/Mg removal–concentration for Li precipitation–three-stage counter-current washing. Optimizing roasting temperature (600 °C), hydrochloric acid concentration (18 wt%), and leaching parameters achieved a 92.37% Li leaching rate. Multi-step purification yielded lithium carbonate with >99% Li₂CO₃ purity and an overall Li recovery of 73.89%. Process II follows blank roasting–sulfuric acid leaching–Al removal via alum precipitation–Al/Fe removal–freeze crystallization for sodium sulfate removal–Ca/Mg removal–concentration for Li precipitation–three-stage counter-current washing. Parameter optimization and freezing impurity removal achieved an 89.11% Li leaching rate, producing lithium carbonate with >98.85% Li₂CO₃ content alongside by-products like crude sodium chloride and ammonium alum. Both processes enable resource utilization of Al-rich residues, with the hydrochloric acid-based method excelling in stability and the sulfuric acid-based approach offering superior by-product valorization potential. This low-energy, high-yield clean extraction system provides critical theoretical and technical foundations for scaling clay-type Li deposit utilization, advancing green Li extraction and industrial chain development. Full article

Wellhead choke performance is critical for flowback choke-size managements in unconventional gas wells. Most existing empirical correlations were originally developed for oil and gas flow, and their accuracy for gas/water multiphase flowback remains uncertain. This study presents a data-driven approach to examine the choke–performance relationship during multiphase flowback. We compiled a flowback dataset containing 18,660 surface measurements from 37 shale gas wells in the Horn River Basin. Using machine learning, we modeled choke performance based on flowback features including water rate, gas/water ratio, wellhead and separator pressures and temperatures, and choke size. The models achieved strong predictive accuracy. Based on the machine learning results, we developed a new choke–performance correlation tailored to multiphase flowback. This model was validated against field data and showed reliable performance. The findings provide a useful tool for optimizing choke-size strategies during flowback in hydraulically fractured gas wells, especially in unconventional reservoirs. Full article

This study presents a numerical investigation into the heart valve through a fluid–structure interaction (FSI) framework using a two-dimensional, steady-state, Newtonian flow assumption. While simplified, this approach captures core biomechanical effects and provides a baseline for future extension toward non-Newtonian, pulsatile, and three-dimensional models. The analysis focuses on the influence of magnetic field intensity characterized by the Hartmann number (Ha) and flow regime defined by the Reynolds number (Re) on critical hemodynamic parameters, including wall shear stress (WSS), velocity profiles, and pressure gradients in the valve region. The results demonstrate that stronger magnetic fields significantly stabilize intravalvular flow by suppressing recirculation zones and reducing flow separation distal to valve constrictions, offering protective hemodynamic benefits and serving as a non-invasive method to modulate vascular behavior and reduce the risk of cardiovascular pathologies such as atherosclerosis and hypertension. Full article

This study presents a novel bivariate-output deep learning framework based on LeNet-5 for the simultaneous prediction of urban flood susceptibility and inundation depth in Seoul, South Korea. Unlike previous studies that relied on single-output models, the proposed approach jointly learns classification and regression targets through a shared feature extraction structure, enhancing consistency and generalization. Among six tested architectures, the Le5SD_CBAM model—integrating a Convolutional Block Attention Module (CBAM)—achieved the best performance, with 83% accuracy, an Area Under the ROC Curve (AUC) of 0.91 for flood susceptibility classification, and a mean absolute error (MAE) of 0.12 m and root mean squared error (RMSE) of 0.18 m for depth estimation. The model’s spatial predictions aligned well with hydrological principles and past flood records, accurately identifying low-lying flood-prone zones and capturing localized inundation patterns influenced by infrastructure and micro-topography. Importantly, it detected spatial mismatches between susceptibility and depth, demonstrating the benefit of joint modeling. Variable importance analysis highlighted elevation as the dominant predictor, while distances to roads, rivers, and drainage systems were also key contributors. In contrast, secondary terrain attributes had limited influence, indicating that urban infrastructure has significantly altered natural flood flow dynamics. Although the model lacks dynamic forcings such as rainfall and upstream inflows, it remains a valuable tool for flood risk mapping in data-scarce settings. The bivariate-output framework improves computational efficiency and internal coherence compared to separate single-task models, supporting its integration into urban flood management and planning systems. Full article

Understanding asymmetric interactions among heterogeneous traffic participants is essential for managing congestion and enhancing safety at urban signalized intersections. This study proposes a cellular automaton modeling framework that captures the spatial and behavioral asymmetries among vehicles, bicycles, and pedestrians, with a particular focus on right-of-way hierarchies and conflict anticipation. Beyond simulation, the framework integrates a behavior pattern mining module that applies unsupervised trajectory clustering to identify recurrent interaction patterns emerging from mixed traffic flows. Simulation experiments are conducted under varying demand levels to investigate the propagation of congestion and the structural distribution of conflicts. The results reveal distinct asymmetric behavior patterns, such as right-turn vehicle blockage, non-lane-based bicycle overtaking, and pedestrian-induced disruptions. These patterns provide interpretable insights into the spatiotemporal dynamics of intersection performance and offer a data-driven foundation for optimizing signal control and multimodal traffic flow separation. The proposed framework demonstrates the value of combining microscopic modeling with data mining techniques to uncover latent structures in complex urban traffic systems. Full article

This study introduces an amphibious spherical robot equipped with a dual-propulsion system (ASR-DPS) and investigates its water-surface motion characteristics. Due to its distinctive spherical geometry, the robot exhibits markedly different hydrodynamic behavior compared to conventional vessels. A comparative analysis of the frontal wetted area is performed, followed by computational fluid dynamics (CFD) simulations to assess water-surface performance. The results indicate that the hemispherical bow increases hydrodynamic resistance and generates large-scale vortex structures as a consequence of intensified flow separation. Although the resistance is higher than that of traditional hulls, the robot’s greater draft and dual-propulsion configuration enhance stability and maneuverability during surface operations. To validate real-world performance, standard maneuvering tests, including circle and zig-zag maneuvers, are conducted to evaluate the effectiveness of the propeller-based propulsion system. The robot achieves a maximum surface speed of 1.2 m/s and a zero turning radius, with a peak yaw rate of 0.54 rad/s under differential thrust. Additionally, experiments on the pendulum-based propulsion system demonstrate a maximum speed of 0.239 m/s with significantly lower energy consumption (220.6 Wh at 60% throttle). A four-degree-of-freedom kinematic and dynamic model is formulated to describe the water-surface motion. To address model uncertainties and external disturbances, two control strategies are proposed: one employing model simplification and the other adaptive control. Simulation results confirm that the adaptive sliding mode controller provides precise surge speed tracking and smooth yaw regulation with near-zero steady-state error, exhibiting superior robustness and reduced chattering compared to the baseline controller. Full article

Synchronous grouting slurry is widely used in shield tunnel construction to fill the gaps between stratum and shield tail segments. However, as grout is nearly liquid in the initial stages, the tunnel lining segments recently separated from the shield tail are easily affected by the upward buoyancy generated by grout, causing issues such as longitudinal misalignment and opening of ring joints. Therefore, studying the upward buoyancy characteristics of synchronous grout is crucial. In this study, floating characterisation parameters of grout were investigated using buoyancy model tests, orthogonal tests, and comprehensive tests. The floating characterisation parameters are affected by distribution ratio and types of each grout component. The relationship between the floating characterisation parameters of grout and buoyancy was established. The results show that density, flow index, and shear strength can be used as the floating characterisation parameters. Binder–sand and water–binder ratios have the largest impact on the density. The bentonite–water ratio exerts a primary influence on the flow index, while the water–binder ratio contributes a secondary effect. In addition, bentonite–water and binder–sand ratios have the greatest effect on the shear strength. Furthermore, the particle size of sand and type of bentonite considerably influence the flow index and shear strength. A high-shear grout using well-graded fine sand and a high mesh of sodium bentonite was considered in this study. When the content of bentonite exceeds 7% (P2.2), Archimedes’ law is not applicable for calculating the upward buoyancy of grout. Buoyancy supply rate exhibits gradual enhancement with flow index elevation, yet with diminishing growth rates. Full article

This study presents an automated industrial wastewater treatment system based on Siemens programmable logic controller (PLC) that optimizes reagent dosing, aeration, sedimentation, and sludge separation. The system uses accurate pH sensors, dosing pumps, solenoid valves, and a human–machine interface (HMI), and real-time monitoring is provided by a Teltonika TRB255 communication module (<45 sec. response time). As a result, the treatment cycle time was reduced by 31%, reagent consumption by 30%, and operator intervention was reduced from 95 to less than 15 min per day, achieving a pollutant removal efficiency of 89%. A two-layer LSTM architecture developed on the PyTorch platform predicts pH (6.7–7.7), temperature (12–20 °C), and reagent consumption (~9.8 kg/cycle). The model was trained with 240 h of data (64 neurons, learning rate 0.001). The validation loss remained stable, indicating reliable learning. The study confirms that AI-based automation provides greater process stability, meets environmental standards, and promotes sustainable resource use. The scientific novelty of this study is the application of an advanced long short-term memory (LSTM) model to predict wastewater treatment process parameters, allowing for accurate prediction of pH, temperature, flow, and reagent consumption, etc. This provides an opportunity to optimize the process and reduce costs, while ensuring high treatment efficiency and stability. Although there are several publications on the application of artificial intelligence models in the field of industrial wastewater treatment, this is a relatively new field, and there are little data in the scientific literature. Full article

To enhance the aerodynamic performance of vertical axis wind turbines (VAWTs) under complex gust conditions, the design parameters of the flap were optimized using the computational fluid dynamics (CFD) method combined with orthogonal experimental design and the SHERPA algorithm, and two gust models with mainly high and low wind speeds were generated by a self-compiling program to investigate the effects of three combinations of the chordwise mounting position of the flap, the moment of inertia, and the maximum deflection angle on the aerodynamic performance of the vertical axis wind turbine. The results demonstrated that adaptive flaps reduced the flow separation region and suppressed the formation and development of separation vortices, thereby enhancing aerodynamic performance. The adaptive flap was found to be more effective in high-speed gust environments than in low-speed ones. The optimal configuration—chordwise position at 0.4C, moment of inertia at 6.12 × 10⁻⁵ kg·m², and a maximum deflection angle of 40°—led to a 57.24% improvement relative to the original airfoil. Full article

Scramjet operation requires a comprehensive understanding of the internal flowfield, encompassing fuel–air mixing and combustion. This study investigates transient flow and flame development within a HIFiRE-2 scramjet engine combustor, which features two opposing cavities and dual sets of fuel injectors—the upstream (primary) and downstream (secondary) injectors. These cavities function as flameholders, creating circulating flows with elevated temperatures and pressures. Shock waves form both ahead of fuel plumes and at the diverging and converging sections of the flowpath. Special attention is given to the interactions among these shock waves and the shear layers along the supersonic core flow as the system progresses towards a quasi-steady state. Driven by increased backpressure, bow shocks and disturbances induced by the normal, secondary fuel injection and the inclined, primary fuel injection move upstream, amplifying the cavity pressure. These shocks generate adverse pressure gradients, causing near-wall flow separation adjacent to both injector sets, which enhances the penetration and dispersion of fuel plumes. Once a quasi-steady state is achieved, a feedback loop is established between dynamic wave motions and combustion processes, resulting in sustained entrainment of reactive mixtures into the cavities. This mechanism facilitates stable combustion in the cavities and near-wall separation zones. Full article