Applied Sciences doi: 10.3390/app16115441

Authors: Maozhu Peng Zhongkai Huang Yiqun Wu Wei Zhang

This paper proves that existing implicit integration schemes for the modified cam-clay (MCC) cannot handle the dependency of critical state stress ratio on Lode angle, because the coaxiality between the deviatoric strain rate tensor and the deviatoric stress tensor, on which these algorithms are built, no longer holds when the Lode angle effect is considered. A more appropriate algorithm is proposed based on closet point return mapping, and a consistent tangent modulus is calculated. After detailed mathematical derivations, the proposed method is examined via four computational examples. The first example includes a series of numerical triaxial tests, for demonstrating the appropriateness of the employed constitutive equations. The second and third examples are convergence tests at the material (integration point) level and finite element (FE) level, respectively. The resulting quadratic converging rate proves that the Jacobian matrix for return mapping and the consistent tangent modulus for finite element implementation are correctly computed. The last example concerns drained penetration of a surface footing. The proposed method is demonstrated to be more efficient and robust in this case than the Abaqus built-in MCC model, which aborted halfway when simulating the same problem.

Applied Sciences doi: 10.3390/app16115440

Authors: Diego Hernán Villarejo-García Aaron Gómez-Parra José Pino-Ortega Adrián Moreno-Villanueva Carlos Navarro-Martínez Josep Pastor Cano

Breast cancer survival has increased substantially in recent years, leading to a growing need to address long-term physical sequelae and functional impairments. Strength training is recognized as an effective intervention for improving physical function in this population; however, the influence of recovery phase on training outcomes re-mains unclear. This study aimed to explore the longitudinal changes in handgrip strength following an 8-week supervised strength training program and to determine whether these observed variations follow an exploratory association according to the time elapsed since surgery. A quasi-experimental, longitudinal, and prospective design was conducted with 30 breast cancer survivors stratified into three recovery phases (≤6 months, 7–12 months, ≥13 months). Participants completed an 8-week supervised strength training program, and handgrip strength was assessed before and after the intervention. A linear mixed model was used to analyze the effects of time, hand condition, and recovery phase using maximum trial values. Significant improvements in handgrip strength were observed across all cohorts, with greater absolute adaptations in the late recovery phase (+3.80 kg, p < 0.001) compared to the intermediate (+2.87 kg, p < 0.001) and early (+2.25 kg, p = 0.015) phases. A persistent inter-limb strength deficit was present on the operated side. These preliminary findings suggest that resistance training induces strength modifications across all recovery strata, though absolute adaptations vary by time elapsed since surgery.

Applied Sciences doi: 10.3390/app16115438

Authors: Zeyad Tareq Ahmed Ercan Aykut

This research presents an integrated intelligent framework for the early detection of wind turbine failures by combining physical modeling with artificial intelligence techniques, with the aim of improving system reliability and reducing maintenance costs. The significance of this work lies in addressing the challenges associated with a lack of real-world data and the limited ability of traditional models to generalize in complex operating environments. To achieve this, a dynamic model was developed using MATLAB/Simulink to generate signals representing thermal and mechanical behavior under various operating conditions. Several AI models were then applied, including SVM, ANN, Autoencoder, and the hybrid CNN–LSTM model. The results demonstrated the superiority of the CNN–LSTM hybrid model in terms of accuracy, achieving an accuracy of 99.84% with a recall value of 1.0000, reflecting its high ability to detect all failure cases. However, the results showed a relative decrease in precision, indicating the presence of some false alarms. This research provides a simulation-based framework that can support predictive maintenance research. However, further validation using real-world data is required to confirm its applicability in practical wind turbine systems.

Applied Sciences doi: 10.3390/app16115439

Authors: Adnan Mujezinović Ajdin Alihodžić Maja Muftić Dedović Nedis Dautbašić

This paper investigates characterization of low-frequency electromagnetic transients occurring in power networks with the Stockwell transform (S-transform). Three characteristic cases of real network events are considered. The experiments are conducted in the 35 kV network, and voltage and current signals are recorded. The S-transform is applied to analyze voltage and current signals recorded during earth fault, ferroresonance and transformer inrush. The obtained results demonstrate that application of the S-transform enables identifying and distinguishing different low-frequency electromagnetic transients. The S-transform scalograms and energy spectrum diagrams provide necessary information to enable more efficient protection measures against problems in the network caused by different low-frequency electromagnetic transient phenomena.

Applied Sciences doi: 10.3390/app16115436

Authors: A. A. Apostolakopoulos P. P. Filippatos K. Davazoglou M. Vasilopoulou C. A. Londos A. Chroneos

Irradiation can impact the properties of semiconductor materials and the function of microelectronic devices. In the present review, we consider how irradiation interacts with semiconductor materials important, primarily silicon (Si), focusing on the defect processes. These, in turn, will have an impact on the physical properties of the material and can impact important properties for devices such as the electrical conductivity and mechanical integrity. We consider the ways that irradiation impacts the operation of microelectronic devices. We thereafter review the defect engineering strategies and other ways to mitigate against the impact of irradiation in devices. Finally, we consider the potentially important role of irradiation defects as qubits in the emerging quantum technologies.

Applied Sciences doi: 10.3390/app16115437

Authors: Abel I. Barrial-Lujan María del Mar Camacho Nuria Martínez-Navarrete Eva García-Martínez

The valorization of agri-food by-products as functional ingredients requires understanding how processing and formulation affect their nutritional and metabolic properties. This study evaluated the combined effects of drying method (hot air drying, HAD; freeze-drying, FD), particle size (80 and 500 µm), and gum Arabic (GA) addition on the compositional and metabolic functionality of faba bean (Vicia faba L.) pod flour. Proximate composition, total phenolic content (TPC), estimated glycemic index (eGI), glucose dialysis retardation index (GDRI), and enzyme inhibitory activities (α-glucosidase and pancreatic lipase) were determined. Results showed that all factors significantly affected eGI, with independent contributions, whereas GDRI was mainly influenced by particle size and GA, with significant interaction effects. GA addition consistently reduced eGI and increased GDRI, indicating improved modulation of both starch hydrolysis and glucose diffusion. HAD samples showed higher enzyme inhibitory activity, while FD combined with GA enhanced TPC. Particle size modulated structural properties affecting starch accessibility and glucose diffusion. Soluble dietary fiber and phenolic compounds were key contributors to in vitro metabolic functionality, while matrix structure determined their effectiveness. These results suggest that faba bean pod powders may serve as sustainable functional ingredients for food applications, contributing to the valorization of agri-food by-products within a circular economy approach.

Applied Sciences doi: 10.3390/app16115435

Authors: Betül Ay Yunus Emre Demirdağ

Serving LLMs to many concurrent users gives rise to significant challenges for latency, throughput, and scalability. This paper presents a systematic and reproducible benchmark of two widely used large language model (LLM) serving frameworks, Ollama and vLLM, under concurrent workloads. We conducted a controlled comparison by running both frameworks on a single NVIDIA H100 80 GB GPU with the same model weights (Qwen3-4B) and inference configurations. We then evaluated them using five open benchmark datasets across four scenarios consisting of baseline question answering, complex reasoning, streaming interaction, and stress testing. Each scenario was executed under increasing concurrency levels, ranging from light loads to high-concurrency stress, to measure end-to-end latency, throughput, time-to-first-token (TTFT), success rate, and resource usage. Our experiments show that vLLM clearly outperforms Ollama across all four scenarios, achieving a 100% request success rate. It delivers 20–29 times higher throughput and 8–19 times lower P95 latency, and it completes every request successfully. Additionally, vLLM produced the first token within 0.5–3.5 s and remained stable at up to 100 concurrent users. Ollama, by contrast, required 54–122 s for the first token, hit a concurrency bottleneck near 10 users, and exhibited a timeout-based error rate of 13–30.06% under heavier loads. Notably, both frameworks demonstrated significant memory growth during extended endurance tests, necessitating careful monitoring in long-running deployments. Overall, for a single model on a single GPU, vLLM is highly suitable for high-concurrency serving, while Ollama remains a practical choice for lightweight, local, or developmental workflows.

Applied Sciences doi: 10.3390/app16115434

Authors: Dongin Park Taehoon Kim

The Itaewon tragedy in South Korea highlighted the severe risks associated with ultra-high-density pedestrian environments. In this study, pedestrian safety in narrow urban alleys was quantitatively evaluated using a FLEXSIM-based Multi-Agent System (MAS) simulation that models individual pedestrian interactions under extremely crowded conditions. Two simulation scenarios were established: a typical alley configuration and a bottleneck condition caused by illegal construction. In addition, three pedestrian control strategies (i.e., bidirectional flow, right-side walking enforcement, and one-way traffic control) were comparatively analyzed. Evacuation time, pedestrian collision frequency, and associated risk levels (Level 0–Level 4) were evaluated according to pedestrian density and movement direction. The simulation results show that bottleneck conditions significantly increase pedestrian collision frequency and evacuation time under high-density conditions. Among the examined strategies, one-way traffic control most effectively reduced pedestrian interactions and evacuation delays, whereas the bottleneck scenario under bidirectional pedestrian flow showed the highest risk level. These findings highlight the importance of pedestrian flow control and bottleneck management in reducing crowd risk in ultra-high-density pedestrian environments and provide quantitative data for pedestrian safety assessment and crowd management planning. Furthermore, the present study provides a quantitative simulation-based approach for analyzing pedestrian collision risk and evacuation safety under ultra-high-density bottleneck conditions in narrow urban alley environments.

Applied Sciences doi: 10.3390/app16115422

Authors: Menghan Wu Shengbo Chen Xitong Xu Yaqi Zhang Yuqiao Suo Jiaqi Yang Jinchen Zhu Aonan Zhang Qiqi Li

Ground control point (GCP) selection is a critical step in the automated high-precision geometric correction of remote sensing imagery. While the quantity, quality, and distribution of GCPs are three factors which may affect the accuracy of geometric correction, traditional automated selection methods predominantly focus on optimizing spatial distribution, often neglecting the inherent quality heterogeneity within matched point sets. This paper proposes a Reliability-weighted Spatial Coverage Sampling (SCS+R) method, which integrates matching reliability into the spatial coverage sampling framework via an adaptive weight factor (α). Experiments using Gaofen-2 (GF-2) imagery demonstrate that with 58 GCPs selected by SCS+R, the relative geometric consistency with the reference imagery is improved to a sub-pixel level (1.55–2.23 m) for multispectral images and within two pixels (0.99–1.81 m) for panchromatic images. Compared to the standard SCS, Voronoi, and weighted Voronoi methods, SCS+R improves the average accuracy by approximately 25%, 16%, and 8%, respectively. These results verify the enhanced stability and robustness of the proposed method in complex environments. Moreover, the optimal adaptive reliability weight α consistently stabilizes in a low range of 0.1–0.3, quantitatively revealing a key principle for small-sample GCP selection: spatial uniformity provides the foundation, while point reliability is the key to achieving high precision.

Applied Sciences doi: 10.3390/app16115430

Authors: Paolo Costa Angela Conti Viviana Paulon Laura Corte Gianluigi Cardinali Sergio Casella Christian Kennes Maria Carmen Veiga Marina Basaglia Lorenzo Favaro

Acidogenic fermentation of organic wastes represents a strategic platform for the co-production of H2, CO2, and volatile fatty acids (VFAs), which are potential key intermediates for cost-effective polyhydroxyalkanoate (PHAs) biosynthesis. This typically relies on carbon sources that are too expensive and hinder the commercialization of PHAs. This study provides metagenomic insights into the microbial dynamics underpinning the acidogenic conversion of waste melon under increasing organic loading rates (OLRs). Metabarcoding revealed that Megasphaera dominated the community, with its abundance rising markedly from 5 to 20 gCOD/L, accompanied by relevant contributions from Solobacterium, Prevotella, and Clostridium. These taxa were associated with the formation of acetic, propionic, and butyric acids and with enhanced hydrogenogenesis. Higher OLRs, up to 20 gCOD/L, promoted hydrogen-producing species while suppressing lactic acid bacteria, thereby improving H2 and VFAs yields up to 26.7% v/v and 13 gCOD/L, respectively. By linking microbial shifts to metabolic outputs, this work advances the understanding of acidogenic pathways essential for integrating dark fermentation-derived H2, CO2, and VFAs into sustainable PHAs production systems.

Applied Sciences doi: 10.3390/app16115433

Authors: Kun Tian Laibin Zhang Shunyi Wang Jinjiang Wang Yujie Cheng

The petroleum and petrochemical industry involves high-risk operations, where traditional manual supervision suffers from blind spots and incomplete coverage, while massive video data remain underutilized. This study collected 1.4 million images from high-risk operation sites and constructed a multi-mechanism hazard identification method using computer vision, integrating object detection, pose estimation, and object tracking. Spatiotemporal attention mechanisms were incorporated to enhance recognition accuracy for multi-scale and small targets. Based on violation behaviors, an algorithmic reasoning logic was designed to automatically identify key targets from complex video images. The study developed 40 video recognition algorithms for operational hazards (e.g., personnel standing under a crane boom and working at heights without a safety harness), achieving an accuracy of ≥90%. These algorithms enable real-time, intelligent identification of violation behaviors, facilitating the transformation of risk management from “human-based defense” to an integrated “human + technical + intelligent defense” model, allowing early intervention and elevating safety risk management standards.

Applied Sciences doi: 10.3390/app16115432

Authors: Dimitra Korda Maria Papandreou Nikolaos Chrysagis George A. Koumantakis

This pilot randomized controlled trial (RCT) aimed to evaluate the effect of a lumbar spine stabilization exercise program (LSSEP) in healthy adults with a sedentary lifestyle. Thirty-eight healthy Greek adults (mean age 40.53 ± 11.91 years) who had been sedentary for at least six months participated in the study. They were randomly assigned to an intervention group (n = 18) and a control group (n = 20). The intervention group followed a supervised progressive four-week LSSEP, while the control group received a booklet with ergonomic and stretching instructions (passive control). The primary outcome was the 12-Item Short Form Health Survey version 2 (SF-12v2), which measures quality of life, and secondary outcomes were the Functional Movement Screen (FMS) and the Y-Balance test (YBT), assessed before and immediately after the intervention period. A repeated-measures 2 × 2 mixed ANOVA revealed significant time × group interactions in favor of the intervention group for the mental component of quality of life and for functional movement and dynamic balance performance (p < 0.05), with moderate to large effect sizes. The LSSEP appears to improve specific aspects of quality of life and movement performance in healthy sedentary adults, with potentially clinically meaningful benefits observed even within a short four-week period. Confirmation through larger, longer-term trials is warranted.

Applied Sciences doi: 10.3390/app16115431

Authors: Mehariw Zewdie Muche Wallelign Mulugeta Nebiyu Ephrem Melaku Getachew Worku Tilahun Tsega Mitiku Damtie Yehualaw Woubishet Zewdu Taffese

Cement is widely used worldwide but contributes to environmental issues due to its reliance on non-renewable resources and high CO2 emissions. Incorporating waste materials, such as eggshell ash (ESA) and granite waste powder (GWP), as partial cement replacements offers a sustainable approach to reducing the environmental impact of mortar production. This study investigated the effects of replacing cement with a blended eggshell ash–granite waste powder (ESAGWP) mixture at 0%, 5%, 10%, 15%, 20%, 25%, and 30% by weight. Experimental tests evaluated fresh, hardened, and microstructural properties, including workability, compressive strength, bulk density, water absorption, porosity, ultrasonic pulse velocity (UPV), and resistance to sulfate attack at curing ages of 3, 7, 28, 56, and 91 days. The results showed that a 15% replacement of cement with ESAGWP provided optimal performance, maximizing compressive strength, bulk density, and UPV, particularly at later curing ages. At this optimal level, compressive strength reached 35.00 MPa, 36.77 MPa, and 37.58 MPa at 28, 56, and 91 days, respectively, representing improvements of approximately 28.0%, 28.8%, and 26.6% over the plain cement control mix at the corresponding ages. Replacements beyond 15% led to reduced strength, increased porosity, and higher water absorption due to unreacted particles. Microstructural analysis revealed that the ESAGWP15 mix achieved a dense and well-packed matrix, correlating with improved mechanical and durability properties. Overall, the study demonstrates that ESAGWP can serve as an effective supplementary cementitious material (SCM), with 15% replacement recommended for balanced performance and sustainability in mortar production.

Applied Sciences doi: 10.3390/app16115426

Authors: Yuliang Jin Duanyang Li Zhiwu Li Naiqi Wu

Traffic accident prediction is a key challenge in road safety, and it is necessary to accurately identify high-risk sections from different data sources. Although graphical neural networks (GNNs) simulate the road network topology well, they ignore the visual and environmental clues from physical road conditions. This paper addresses this gap by proposing a Sequential Geometric Reasoning Network (SGR-Net), a deep learning framework for multimodal accident prediction. Unlike prior GNN-based approaches, SGR-Net introduces a Geometry–State Attentive Fusion (GSAF) module—its main novelty—which dynamically integrates visual features from satellite imagery with structural graph contexts. The framework also includes a stability-aware training objective and meta-learning for cross-region generalization. We evaluate on a large-scale dataset covering six U.S. states with over nine million accidents and one million satellite images. SGR-Net achieves strong results, with AUROC up to 96.8% and MAE as low as 0.08 in Delaware. Ablations confirm the GSAF module is essential: removing it reduces AUROC by 2.7% and increases MAE by over 40%. The framework establishes a new state-of-the-art for multimodal traffic accident prediction.

Applied Sciences doi: 10.3390/app16115429

Authors: Dariusz Janiszewski

This paper presents the design and implementation of a field-programmable gate array (FPGA)-based System-on-Programmable-Chip (SoPC) architecture for sensorless speed control of permanent magnet synchronous motor (PMSM) drives. To enable real-time execution of the computationally intensive estimation stage, a parallelized Unscented Kalman Filter (UKF) is proposed for the joint estimation of rotor speed, position, and load torque. Unlike traditional sequential processor-based UKF implementations, the proposed parallel architecture simplifies the iterative process and significantly reduces computational latency and hardware resource utilization while preserving high estimation fidelity. This transformation reduces the number of sequential dependency stages within one estimation cycle and enables simultaneous execution of matrix operations using dedicated FPGA resources, thereby decreasing effective iteration latency. The complete control system comprises current regulators, a coordinate transformation module, a proportional–integral (PI) speed controller, and auxiliary functional blocks—all fully integrated within a single SoPC. The UKF estimator and control components are described using a hardware description language (HDL), enabling efficient hardware-level parallelism and real-time execution. The proposed system is validated through co-simulation and experimental verification on a Xilinx ZCU102 platform driving an inverter-fed PMSM. The results confirm correct real-time operation of the proposed architecture and demonstrate its feasibility for FPGA-based sensorless motor drive implementation. A detailed quantitative comparison with a fully sequential FPGA-based UKF implementation is identified as future work to further substantiate the reported latency reduction.

Applied Sciences doi: 10.3390/app16115428

Authors: Cezary Graul Wojciech Żarski Dariusz Mikołajewski Izabela Rojek

Industry 4.0 and 5.0 technologies have made industrial environments data-rich, yet a persistent cognitive gap remains: operators face substantial difficulty interpreting and acting on this data in unstructured, time-critical situations. This paper presents an architecture that integrates the Asset Administration Shell (AAS), OPC UA, and a Large Language Model (LLM)-based agentic AI within a mandatory Human-in-the-Loop (HITL) framework. The AAS acts as a semantic grounding layer through Retrieval-Augmented Generation (RAG), supplying the LLM agent with ECLASS-referenced technical parameters that reduce the risk of hallucination. OPC UA Methods form a deterministic execution layer that keeps agent actions within PLC-validated safety boundaries. The HITL mechanism enforces a cryptographic approval gate so that no physical machine action can occur without documented human authorization. This requirement was motivated by an industrial survey (n=117), in which 47% of employees stated that human oversight is irreplaceable, combined with enterprise safety and accountability requirements and broader governance considerations for AI-driven actuation in safety-critical cyber-physical systems. Two proof-of-concept case studies evaluate the architecture under controlled laboratory conditions. Proof-of-concept results indicate system processing latencies of 1.7 s (maintenance) and ∼15 s (scheduling), with end-to-end latencies (including mandatory human approval) of 14.9 s and 62 s, respectively, representing estimated improvements of approximately 97% and 96% over expert-estimated manual baselines (∼8 min and 25–40 min). All figures derive from single scripted runs under controlled laboratory conditions and should be read as indicating architectural feasibility at Technology Readiness Level 4, not as statistically validated performance benchmarks: variability bounds and confidence intervals are unavailable, the manual baselines are expert estimates rather than instrumented measurements, and operator deliberation times derive from a single response per scenario. A structured comparison with related work shows that, to the authors’ knowledge, no published approach in the surveyed literature combines AAS semantic grounding, OPC UA deterministic execution, and mandatory cryptographic HITL within a single empirically grounded framework.

Applied Sciences doi: 10.3390/app16115427

Authors: Ewa Radwan Mateusz Kania Karolina Widzisz Joanna Zyla Agnieszka Szczęsna Andrzej Polański

Expectation Maximization (EM) iterations for Gaussian mixture models (GMMs) are highly sensitive to initial parameters, which calls for developing robust initialization methods. For multidimensional GMMs, the problem is even more severe than for univariate GMMs because of the larger number of parameters. For univariate or low-dimensional GMMs, several studies on their initialization have appeared in the literature, whereas research on initializing multivariate, high-dimensional GMMs remains limited. In this study, we compare several initializations for Multivariate Diagonal Gaussian Mixture Models (MDGMMs). In our study, we have included methods already used in the literature for initializing MDGMMs: Hierarchical Clustering, K-means, and random initialization. We have also used a new method, namely, Ensemble Clustering. A review of the existing literature suggests that Ensemble Clustering has not been used previously as an initialization strategy for MDGMM. Several metrics were used to evaluate the clustering quality. Our study demonstrates that Ensemble Clustering, while computationally intensive, is competitive with other methods for initializing MDGMMs.

Applied Sciences doi: 10.3390/app16115425

Authors: Alkistis Kevrekidou Andreana N. Assimopoulou

Cymodocea nodosa, growing at low water depth, is affected by various environmental changes and is expected to adapt to oxidative stress. Oxidative stress in living leaves (LC) and beach deposits (NC) of C. nodosa activated superoxide dismutase (SOD), which was higher in LC, leading to significant neutralization of the produced H2O2 and destruction of protein generation. Higher antioxidant capacity (using a UV/Vis spectrophotometer) to scavenge 2.2-azino-bis(3-ethylbenzthiazoline-6-sulfonic acid (ABTS•+) (IC50: 5 in LC vs. 22 μg mL−1 in NC) and OH• (hydroxyl) radicals (IC50: 132 in LC vs. 281.7 μg mL−1 in NC), compared to 2,2-diphenyl-1-picrylhydrazyl (DPPH•) (IC50: 63 in LC vs. 45 μg mL−1 in NC) and superoxide anion (O2•−) radicals (IC50: 190 in LC vs. 94 μg mL−1 in NC), and similar reducing power (RP) were recorded in LC compared with NC extracts (IC50: 53 in LC vs. 52 μg ml−1 in NC). Phenolic compounds were not significantly lost during plant exposure on shores (mean value: 57.00 in LC vs. 45.48 mg g−1 d.w. in NC). Phenolic compounds identified, using UHPLC-DAD analysis, in both LC and NC extracts were chicoric, trans-ferulic, caftaric, p-coumaric, sinapic, and trans-cinnamic acid and rutin hydrate, whereas caffeic acid, in traces, was identified in NC extracts. NC showed higher cytotoxic activity in inhibiting LS 174 colon cancer cells than LC. In cases of plant cultivation or management plans for seagrass meadows and their beach deposits, with the least possible impacts, both LC and NC extracts could be exploited for their antioxidant and anticancer properties. In a ‘case study’, the amounts of individual phenolic compounds that can be produced from NC utilization were estimated.

Applied Sciences doi: 10.3390/app16115424

Authors: Jie Guo Ying Zhang Yanhan Zhao Yi Cao Kailei Chen Qian Zhou Chao Yuan

This paper proposes an improved ant colony optimization-based path planning method for UAV inspection in substations. Considering the equipment partition characteristics and no-fly zone constraints, a two-dimensional inspection scenario model is constructed with typical equipment areas, inspection points, a depot, and no-fly zones. The fixed partition with the nearest-neighbor method is used as the baseline, and the basic ACO algorithm is introduced for global path search. To further improve path quality, candidate neighborhood selection, elite pheromone updating, integrated turning and obstacle-avoidance costs, and local optimization are incorporated into the improved ACO. Simulation results based on 30 independent runs show that the improved ACO achieves an average path length of 1694.08 m and an average estimated flight time of 372.27 s in the 24-point scenario, reducing these two metrics by 22.30% and 20.89%, respectively, compared with the baseline method. Compared with the basic ACO, the improved ACO further reduces the average path length and estimated flight time by 2.28% and 2.41%, respectively, with statistically significant differences. Comparisons with GA and PSO and scalability experiments under different inspection point scales further demonstrate the effectiveness of the proposed method.

Applied Sciences doi: 10.3390/app16115423

Authors: Javier J. Ruiz Gabriel Mauricio Ramírez Villegas Jaime Díaz-Arancibia Ana Bustamante-Mora

Brain–computer interface (BCI) systems aim to establish direct communication pathways between neural activity and external devices, enabling interaction without relying on conventional neuromuscular mechanisms. This study investigates the feasibility of decoding binary decisions (“Yes”/”No”) from self-generated cognitive–emotional modulation patterns using a single-channel low-cost EEG device. The proposed approach evaluates whether internally generated modulation strategies can produce distinguishable neural activity suitable for BCI applications under constrained acquisition conditions. EEG signals were recorded from two participants using a consumer-grade headset while they responded to questions through intentional internal modulation associated with affirmative and negative responses. The recorded signals were preprocessed, and multiple feature representations were extracted, including raw temporal data, cepstral coefficients, spectral power, and continuous wavelet transform (CWT) features. Several machine learning and deep learning models, including convolutional neural networks (CNN), long short-term memory networks (LSTM), and support vector machines (SVM), were trained and evaluated using hold-out and stratified k-fold validation strategies. The best performance was achieved by a CWT-based CNN model, reaching an average accuracy of 80.5%, significantly above chance level. Additional models, including CEP-CNN and RAW-LSTM, achieved competitive results, highlighting the relevance of feature representation in EEG-based classification tasks. The results demonstrate that internally generated modulation patterns can produce distinguishable EEG responses, even when using low-cost single-channel hardware. Although the limited number of participants constrains statistical generalization, this work serves as a proof-of-concept and provides a reproducible experimental pipeline for future studies. Overall, the findings support the development of accessible, scalable, and user-centered BCI systems based on internally generated neural modulation strategies, contributing to more natural interaction paradigms in EEG-based communication systems.

Applied Sciences doi: 10.3390/app16115421

Authors: Ji-Hyen Soung Soon-Pill Jeong Sung-Hun Kim Seong-Sik Kim Youn-Kyung Choi Yong-Il Kim

This study evaluated the biomechanical effects of different attachment configurations on buccal uprighting of lingually inclined mandibular second molars using clear aligners (CAs). Finite element analysis was performed to simulate a 2° uprighting movement, with boundary conditions applied to the basal bone and loading generated through geometric mismatch between the aligner and dentition. Six experimental groups were established according to attachment placement on the first and second molars. Three-dimensional tooth displacement, crown-to-root displacement ratio (C/R ratio), moment-to-force ratio (M/F ratio), and anchorage loss were analyzed. Group 5, combining a buccal attachment on the first molar and a lingual attachment on the second molar, demonstrated the most favorable biomechanical performance, with controlled tipping (C/R ratio −1.67), minimal anchorage loss (3.2%), and an optimal M/F ratio (5.5). In contrast, Group 3 exhibited excessive anchorage loss (44.8%) and inefficient force transmission, while Group 6 showed reduced efficiency despite a high M/F ratio, indicating mechanical overconstraint. These findings suggest that attachment configuration plays a critical role in determining force–moment systems in CA therapy, and increasing attachment number does not necessarily improve treatment efficiency. Strategic placement of attachments can enhance biomechanical control and optimize clinical outcomes in mandibular second molar uprighting.

Applied Sciences doi: 10.3390/app16115420

Authors: Feng Wang Jiadong He Xinghua Chen Md Masum Mia

To enhance crack detection accuracy during UAV-based inspections and address key challenges such as false positives from complex backgrounds, missed narrow cracks, and insufficient structural continuity modeling, this study proposes CFM-Net, a task-oriented segmentation network integrating Channel-Spatial Attention and Multi-Scale Structural Enhancement. Constructed on an optimized U-Net backbone, it employs three dedicated modules: the Channel and Spatial Attention Module (CBAM) to amplify crack-related features and suppress background interference; the Gated Fusion Module (GFF) to dynamically fuse multi-level features, improving detection of fine, narrow cracks; and the Morphology-Guided Multi-Scale Structural Perception Module (MGMSIB), designed to model the structural continuity and multi-scale characteristics of cracks. Comprehensive evaluations on the Mix Bridge Crack dataset demonstrate CFM-Net achieves competitive performance among the evaluated methods, with an mIoU of 80.05% and an F1-score of 87.06%. This represents a significant improvement over strong baselines, outperforming DeepCrack and CrackFormer by 2.3% and 2.42% in mIoU, and 1.21% and 1.03% in F1-score, respectively. Furthermore, the model demonstrates robust performance on heterogeneous crack datasets composed of multiple public sources, particularly in reducing false alarms, recovering narrow cracks, and maintaining crack topology. These results conclusively validate the effectiveness and practical utility of the proposed method for automated bridge crack inspection.

Applied Sciences doi: 10.3390/app16115417

Authors: Yaobai Sun Kang Yang Qiao Chen Hong Yin Yongchang Liang

In the published version of this article [...]

Applied Sciences doi: 10.3390/app16115416

Authors: Ahmed El-Mallul Małgorzata Katarzyna Kowalska Karolina Sawicka Sara Małgorzata Orłowska Łukasz Bednarczyk Łucja Radziszewska

Breast cancer is one of the most commonly diagnosed malignant tumors worldwide and represents a significant public health problem. This paper presents the characteristics of the disease, with particular emphasis on risk factors, mechanisms of development, and molecular classification. Current diagnostic methods and available therapeutic strategies, such as surgery, chemotherapy (CT), radiotherapy (RT), and targeted therapies, are discussed. Particular attention is given to nanotechnology as a promising direction for the development of modern medicine. The potential applications of nanoparticles (NPs) in the diagnosis and treatment of breast cancer are presented, taking into account their mechanisms of action, potential clinical benefits, and limitations related to safety and efficacy. NPs may contribute to increased diagnostic precision and therapeutic efficacy, indicating their significant potential in the future of oncology.

Applied Sciences doi: 10.3390/app16115419

Authors: Mateusz Bielicki Natalia Godlewska Aleksandra Janecka Adrianna Kolas Adrian Topolski

Titanium implants can achieve higher osseointegration when covered with titania nanotubes (TNT). Given their specific morphology, titania nanotubes are excellent substrates for subsequent modifications. In addition to anti-inflammatory drugs, cytotoxic drugs can also be used. It can be achieved by simple physical adsorption of the drug molecules or by their covalent bonding to the surface using a bridging ligand such as (3-mercaptopropyl)trimethoxysilane (MPTMS), for example. The last method was used successfully before. The purpose of the study is to test different modifications of this method to analyze factors that will improve the studied methodology. The study compares two methods of TNTs modification with cisplatin (CDDP) and its oxalate analog (CDOP): drop casting (DC) and the application of MPTMS and its ethoxy analog, MPTES, as bridging ligands. Pluronic L-61 and alkaline Piranha solutions were used as surface activators for TNT. Both activators are effective. Analysis of the fabricated samples was executed using ATR, SEM, SEM/EDX, and AFM. Covalent bonding of Pt(II) complexes to the TNT arrays with a bridging ligand results in a homogeneous layer containing Pt(II) complexes. They release the surface within one hour (the mean values of the kobs for both complexes release in PBS and water are 9 · 10−3 s−1 and 4.8 · 10−3 s−1, respectively). Loading the Pt(II) complexes by drop casting yields layers with higher Pt (II) concentration (ca. 7.5%wt vs. ca. 3.2%wt for the second method and its variants) but lower homogeneity. No distinct general trends in the release rate on the TNT diameter were detected. The results show that modifying Ti6Al4V implants with titania nanotubes and further modifying them with platinum(II) complexes yields materials that can serve as carriers for anticancer platinum-based drugs.

Applied Sciences doi: 10.3390/app16115418

Authors: Chuanhui Gong Qiufeng Wu

Guttation plays an important role in increasing rice yield, preventing crop diseases and improving soil fertility, and is an important index to measure the water status in the field. However, the real-time detection of droplets generated by guttation is a difficult task. Droplets are small and dense targets. In the deep learning model, the detection of small objects in high-resolution images is a key problem to be solved. A Droplet-YOLO method is proposed, which combines the YOLOv8 model and multi-instance learning. The original image is divided into multiple sub-images using a sliding window and detected by YOLOv8. The detection of each sub-image is carried out by an independent repeated test, so multi-instance learning ensures 99.99% probability of detecting droplets. By comparing various models, Droplet-YOLO performs best in terms of precision, recall, and mean precision (mAP). In addition, through the hyper-parameter adjustment experiment, the configuration with a batch size of 32 and epoch of 200 was finally selected. The accuracy and recall rate of the model reached 96.8% and 95.7%, respectively, and the mAP reached 99.0%. Experiments show that this method has significant advantages for small object detection in high-resolution images, and the proposed Droplet-YOLO model is superior to the most advanced methods.

Applied Sciences doi: 10.3390/app16115415

Authors: Tuyu Bao Kun Chen Hao Zhang Zheng Zhang Qingsong Ai Junwei Yan

In the original publication [...]

Applied Sciences doi: 10.3390/app16115414

Authors: Yu Tang Wenbo Luo Jiawei Liu Jingze Yan Fu Xu

Polyethylene (PE) pipes are widely employed in urban gas and water conveyance systems due to their excellent corrosion resistance, cost efficiency, and long service life. However, creep-induced delayed failure remains a critical threat to long-term operational safety and may lead to leakage accidents. Accurate and efficient prediction of creep rupture life is essential for risk control and structural design. This study investigated the performance of four commercial polyethylene pipes, including two PE80-grade and two PE100-grade pipes. By combining the creep test with the critical strain criterion, an efficient and reliable method for predicting the ductile failure lifetime was developed. Creep tests were carried out on dumbbell specimens cut from PE pipes under multiple temperature and stress levels. The time-hardening model was adopted to characterize the nonlinear viscoelastic creep evolution, and the ductile failure time was determined by introducing the critical strain threshold. The predicted lifetimes were systematically validated against experimental data from long-term hydrostatic tests. Results show that the predicted failure times agree well with the measured values, verifying the accuracy and engineering applicability of the proposed method. This approach provides a high-efficiency alternative to conventional long-term hydrostatic tests, offering valuable support for material selection, safety evaluation, and engineering design of PE pipeline systems.

Applied Sciences doi: 10.3390/app16115413

Authors: Raúl Nieto-Acevedo Enrique Alonso-Pérez-Chao Antonio Reyes-Mora Francisco Gallardo Marmol Dimitrije Cabarkapa Jorge Lorenzo Calvo

This systematic review aimed to evaluate the tests used to assess neuromuscular performance in adult female basketball players and to provide evidence-based recommendations for practice and future research. Following PRISMA guidelines and registered in PROSPERO (CRD42025638889), four databases were systematically searched from inception to April 2026. A total of 62 studies were included in the qualitative synthesis and 39 in the quantitative analysis. The most frequently reported assessments examined anthropometry, muscular power, linear speed, change-of-direction (COD) performance, strength, anaerobic capacity, and aerobic capacity. However, substantial variability was observed in testing protocols, outcome variables, and reporting methods. Across studies, performance outcomes showed considerable overlap between competition levels, suggesting that competitive standard alone is not a reliable indicator of neuromuscular performance. Differences in anthropometric characteristics and physical performance were largely influenced by playing position and contextual factors. A key finding was the predominant reliance on outcome-based metrics (e.g., jump height, sprint time), with limited use of force–time variables that provide deeper insight into neuromuscular function. In addition, important methodological limitations were identified, including inconsistent testing procedures, lack of standardized reporting, and the absence of female-specific considerations such as menstrual cycle status. To address these limitations, this review proposes a practical testing framework that integrates reliable, sport-specific, and time-efficient assessment methods. Future research should prioritize the implementation of standardized protocols, the inclusion of force–time analysis, and the development of large-scale descriptive datasets specific to female basketball players. These advances are essential to improve performance monitoring, optimize training prescription, and enhance injury risk management in this population.

Applied Sciences doi: 10.3390/app16115411

Authors: Ioanna Michou Ioannis Maroulis Ioannis Chatzilygeroudis Constantinos Koutsojannis

Patient safety remains a global priority, with surgical errors, including in-hospital infections, procedural mishaps, and delays in diagnosis or treatment, causing over 7 million adverse events and 1 million deaths annually. This study evaluates machine learning (ML) to predict the risk of medical errors in the general surgery department of a Greek tertiary university hospital. Medical error risk was operationalized using several proxy indicators derived from prolonged hospitalization and elevated treatment costs, validated against diagnosis-related complication codes. Using a 10-year dataset of 19,965 anonymized patient records (13.5% error cases, n = 2700), we applied ensemble ML algorithms via WEKA, achieving 94.3% accuracy (Random Forest) in detecting errors such as healthcare-associated infections (HAIs), medication errors, and equipment failures. Given the clinical importance of minimizing missed adverse events, model evaluation prioritized both sensitivity and AUC-ROC in addition to overall accuracy. Key predictors were hospitalization duration (ranked #1 via information gain) and initial diagnosis, enabling early risk flagging (e.g., post-op day 5). Further exploratory analyses excluding hospitalization duration from the predictor set demonstrated a moderate reduction in predictive performance while preserving clinically meaningful discriminative capability, suggesting that model performance was not exclusively dependent on hospitalization duration. Compared to US benchmarks like ACS NSQIP (90% accuracy), our model outperformed by 4.3%, filling a gap in EU/Greek studies amid data silos and resource constraints. Integration with tools like the WHO Surgical Safety Checklist could enable proactive interventions, such as enhanced monitoring for prolonged stays. However, the proposed framework should be interpreted as identifying high adverse-event risk patterns rather than directly detecting clinically adjudicated preventable medical errors. Limitations include retrospective biases and workflow integration challenges; ethical issues like data privacy and algorithmic fairness were addressed via anonymization and ethics approval. Future work will focus on multi-center validation, calibration analysis, longitudinal modeling, and integration of explainable artificial intelligence (XAI) techniques to improve transparency and clinical trust. By blending ML with clinician expertise, this approach shifts healthcare from reactive to proactive error mitigation, improving outcomes and reducing costs.

Applied Sciences doi: 10.3390/app16115412

Authors: Aria Dolatabadian Khalil Kariman

Accurate quantification of soil organic matter (SOM) and carbon content is critical for understanding climate change, evaluating soil health, supporting agricultural sustainability, and implementing carbon sequestration policies. For decades, classical analytical and statistical approaches have underpinned soil carbon assessment, but the emergence of machine learning (ML) techniques offers new opportunities to improve prediction accuracy, scalability, and efficiency. This review summarises the current knowledge on classical and ML-based approaches for analysing SOM and carbon content. We examine the strengths, limitations, and practical applications of conventional methods, including wet chemistry, dry combustion analysis, and geostatistical techniques, alongside modern ML approaches such as random forests (RFs), gradient boosting machines, neural networks, deep learning, and hybrid ML-geostatistical frameworks. Special emphasis is placed on comparative analysis across dimensions, including prediction accuracy, computational requirements, data availability needs, interpretability, uncertainty quantification, and scalability. Soil carbon stocks and dynamics are tightly regulated by indigenous soil microbial communities and their management-driven alterations, creating substantial biologically driven variation that remains difficult to capture with current modelling approaches. We therefore explore hybrid approaches that integrate classical pedological knowledge with ML capabilities. Finally, we discuss emerging challenges, future research directions, and the complementary role these approaches play in advancing soil carbon science. This review concludes that neither classical nor ML approaches alone are sufficient for accurate carbon assessment across diverse scales and environments. Instead, their strategic integration, combining classical mechanistic grounding alongside machine learning’s scalability, represents the most promising path toward realistic soil carbon evaluation for climate change mitigation and agricultural sustainability.

Applied Sciences doi: 10.3390/app16115410

Authors: Le Lei Haigang Ren Xu Li Jianwei Li Boxin Zuo

During aeromagnetic surveys using fixed-wing aircraft, magnetometers mounted inside the cabin are strongly affected by platform magnetic interferences. Existing aeromagnetic compensation models have difficulty in accurately modeling these complex magnetic interferences and often suffer from limited generalization capability. This paper proposes an intelligent compensation algorithm integrating a dual-fluxgate physical extension with an enhanced attention mechanism. First, an extended Tolles–Lawson (T-L) model leverages physical complementarity between dual fluxgates to enhance interference feature representation. Building on this, a physics-informed dual-branch parallel network architecture is designed. The physical branch dynamically models and decouples linear interference components, while the nonlinear branch introduces an improved attention mechanism to capture and remove non-stationary nonlinear interference in the signals. More importantly, the dual-branch network demonstrates superior generalization in level-flight extrapolation tests. Compared to traditional linear methods and pure data-driven models, the proposed approach reduces the residual standard deviation (STD) to 0.16 nT and achieves an improvement ratio (IR) of 22.31. This research significantly advances aeromagnetic compensation precision, offering a robust and high-performance solution for precision geophysical exploration.

Applied Sciences doi: 10.3390/app16115409

Authors: Xueqiao Wang Chenhao Huang Xingkui Fan

Quantum error correction (QEC) is indispensable for suppressing noise in near-term and future quantum processors. Most neural decoders proposed for the surface code exploit predominantly local syndrome neighborhoods, which limits their ability to capture lattice-wide correlations. To overcome this limitation, we develop a two-stage learning-based decoding framework that leverages transformer self-attention to provide a global receptive field. In the first stage, a low-level decoder (LLD) based on a feedforward neural network predicts physical corrections from measured syndromes. In the second stage, a high-level decoder (HLD) performs logical-level verification by classifying the logical equivalence class and checking logical consistency; the HLD is instantiated using either a convolutional network or a transformer encoder. Monte Carlo experiments under depolarizing noise demonstrate clear threshold gains: for code distance d = 3, 5 and 7, the threshold improves from 0.03339 ± 0.00072 (LLD only) to 0.04110 ± 0.00085 with a CNN-based HLD, and further to 0.05350 ± 0.00112 when the HLD is implemented with a transformer; comparable improvements are observed for larger code distances. These results indicate that explicit logical-level discrimination mitigates decoding failures caused by degeneracy, and that global attention better captures long-range topological structure than convolutional baselines.

Applied Sciences doi: 10.3390/app16115408

Authors: Qian Wang Yini Zhou Haowen Shi Qian Huang

Human action recognition is a key research area in computer vision, where accurate recognition relies on effective modeling of both global and local spatiotemporal information. However, existing GCN-based methods often overemphasize the local topological connectivity of human skeletons. Moreover, their temporal modules fail to fully capture the evolution of action sequences, leading to critical instantaneous information being obscured by global representations. To address these problems, we propose an integrated framework termed MADS-GCN. In the spatial modeling stage, we introduce two parallel streams: the Physical Stream uses the adjacency matrix to constrain convolution and capture global structural patterns, while the Topological Stream leverages spatial attention to assign adaptive weights to joints, preserving discriminative local adaptive features. For temporal modeling, a channel-temporal attention mechanism is applied to adaptively refine feature maps, followed by a bidirectional GRU to capture multi-scale temporal patterns. Extensive experiments on NTU RGB+D60, Northwestern-UCLA, and our custom DanceBasic-Set demonstrate the effectiveness of MADS-GCN and indicate its applicability to dance action recognition scenarios.

Applied Sciences doi: 10.3390/app16115407

Authors: Aliona Ghendov-Mosanu Iurie Rumeus Sorina Ropciuc Olesea Saitan Viorica Bulgaru Svetlana Leatamborg Galina Lupascu Georgiana Gabriela Codină

Triticale grains and brewers’ spent grain (BSG) offer promising, sustainable ingredients for bread development, as triticale adapts well to climate change and BSG is a low-cost by-product supporting zero-waste goals. This study evaluated the rheological properties of dough and bread quality obtained from seven triticale cultivars (Ingen 35, Ingen 93, Ingen 40, Ingen 33, Ingen 54, Costel, and Fanica) grown in the Republic of Moldova, with the addition of 5% and 10% fermented dark BSG (BSGF). BSGF incorporation decreased dough stability and protein network strength, as indicated by Mixolab parameters, while the pasting properties varied according to the cultivar. Dynamic rheology showed reductions in storage (G′) and loss (G″) moduli, with tan δ < 1 for all samples. Increasing BSGF levels reduced falling number, Alveograph tenacity, extensibility, baking strength, and Rheofermentometer parameters. In bread, BSGF addition decreased loaf volume and porosity while significantly increasing acidity. Color analysis showed reduced lightness (L*) and increased redness (a*). Texture profile analysis indicated increased hardness and adhesiveness, with stable cohesiveness and reduced resilience. Sensory evaluation revealed improved color and a “hearty” texture at 5% inclusion, whereas 10% resulted in a denser structure and lower acceptability. BSGF significantly influenced the rheological, physicochemical, and sensory properties of triticale bread, highlighting the need for formulation optimization.

Applied Sciences doi: 10.3390/app16115406

Authors: Lintong Li Yunhao Li Washington Yotto Ochieng William Graham Proud Mingyang Huang Mireille El Hajj Arnab Majumdar

Airports, as Critical National Infrastructure (CNI), operate as tightly coupled socio-technical systems exposed to multifaceted threats, including cyber, physical, social, environmental, and Chemical, Biological and Radiological (CBR) threats. This study presents a structured review of the synthesis of conceptual frameworks, airport structural configurations, sensor networks, and multi-domain threat landscapes, as well as airport security and resilience analysis, while comparatively examining risk assessment approaches. The review shows that existing approaches are effective for threat identification and prioritisation but remain predominantly static, with limitations in scalability, data dependency, and real-time applicability. To address these limitations, Threat-Vulnerability-Risk Assessment (TVRA) is adopted as a structured, reusable approach to support metric allocation, redundancy design, and emergency capability development. It further serves as a bridge between traditional risk assessment and resilience-oriented system design by enabling the transformation of static risk scores into scenario-based inputs, thereby supporting stress-testing and lifecycle-based resilience planning across the prepare, act, and recover phases. However, its inherently static structure limits its ability to capture temporal dynamics and cascading interdependencies, highlighting the need to integrate it with dynamic modelling approaches.

Applied Sciences doi: 10.3390/app16115405

Authors: Jakub Komínek Jiřina Száková Lukáš Praus Jiřina Sysalová Luka Stefanović Pavel Tlustoš

The distribution of mercury (Hg) species in soils, as affected by (i) the physicochemical and biological properties of the soils, (ii) the lithogenic vs. anthropogenic sources of soil Hg pollution, and (iii) soil–plant interactions, was investigated in a model pot experiment in which Sinapis alba (Brassicaceae) was planted. The pseudototal (aqua regia-soluble) Hg contents in soils originating from the vicinity of the former cinnabar mine varied between 16.6 and 44.7 mg/kg, whereas in soils from sites where the amalgamation technique had been used for the extraction of gold from ore-bearing materials, the pseudototal Hg values varied between 1.63 and 10.1 mg/kg. However, Hg accessibility was low, with mobilizable Hg(II) accounting for 3–11% of total soil Hg and mobilizable methylmercury (MeHg) remaining below 1%, indicating a limited bioavailable pool under the studied conditions. Mobilizable Hg(II) showed significant negative relationships with total soil carbon and cation exchange capacity (CEC), reflecting its strong association with charged functional groups of the soil sorption complex. The low Hg accessibility in the soil resulted in low Hg contents in plants, not exceeding the feed safety thresholds, with a significant proportion of Hg taken up by the plants being retained in the roots. The results of the determination of gaseous elemental mercury (GEM) indicated its relevance in soil mercury cycling, where further research on the role of plants in GEM emissions is necessary. In this context, the GEM concentrations increased in plants found in soils collected close to the former cinnabar mine. These aspects should be investigated further in future studies.

Applied Sciences doi: 10.3390/app16115404

Authors: Luka Šimić Dario Faj Anja Tomić Ivor Dukić Hrvoje Brkić Turk Tajana Vjekoslav Kopačin

When pregnant patients undergo diagnostic and therapeutic radiological procedures, the unborn child is exposed to an increased risk due to the use of ionizing radiation; therefore, the fetal dose must be estimated and optimized. Tools and methods routinely used for fetal dose estimation lack better personalization of patients. To address this, we developed a software tool for creating phantoms at different pregnancy stages and with varying patient anatomies to further personalize fetal dose estimation using measurements and Monte Carlo simulations. The tool is developed and incorporated into 3DSlicer as a plugin. Phantoms are created based on a real patient phantom, Tena, and physiological data for use in radiological protection. Phantoms are developed with only soft, lung, and bone tissue substitutes, represented for the mother and unborn child. This enables the construction of segmented voxel models as well as mesh models (with the ability to export geometries to DICOM format) of the anatomical structures of pregnant women. Additionally, it allows real patient image registration to enable better personalization of the phantom. The tool can help decrease uncertainty in fetal dose estimation, as well as simplify and accelerate the process of fetal dose estimation. It is released publicly to enable further research.

Applied Sciences doi: 10.3390/app16115403

Authors: Wei Qiao Huan Zhu Peng An Jie An

During the processes of oil extraction, storage, and transportation, pollution caused by frequent fluctuations in the groundwater level and interface changes is complex, and oil leakage has become a major challenge. To effectively identify the water–oil displacement process and the influence of hydrodynamic conditions on displacement efficiency, this paper presents the results of multiple laboratory experiments, focusing on the effects of hydraulic head, medium particle size, displacement direction, and apparatus dimensions on oil displacement efficiency. Three types of media were used: fine sand, medium sand, and coarse sand, simulating two displacement directions—top-down and bottom-up. Displacement efficiency was evaluated by integrating gravity drainage and hydrodynamic displacement processes. The results show that medium particle size has a significant impact on displacement efficiency: coarse sand achieves the highest displacement efficiency (98.74%), whereas fine sand yields the lowest (61.7%). Increasing the hydraulic head significantly improves oil displacement efficiency. The top-down displacement direction exhibits higher efficiency due to gravitational assistance. In addition, apparatus dimensions and sand sample height also significantly affect displacement efficiency. Increasing the apparatus diameter enhances gravity drainage efficiency, while increasing the sand sample height reduces hydrodynamic displacement efficiency. This study provides a scientific basis for remediating LNAPL-contaminated sites and suggests optimizing hydraulic head and displacement direction in practical applications to improve remediation efficiency.

Applied Sciences doi: 10.3390/app16115401

Authors: Zheng Chao Fengfeng Bie Qianqian Li Wensheng Su Tiantian Wei Han Dong

Aiming at solving the problems of the complex impact vibration characteristics of reciprocating compressor valves, the inability of a single signal to fully characterize state characteristics, and the difficulty of effectively extracting and fusing feature information from multi-source signals, this paper constructs a fault diagnosis and prediction model combining Improved Swarm Decomposition (ISWD) and t-SNE dimensionality reduction and fusion with a Multi-scale Convolutional Neural Network–Bidirectional Gated Recurrent Unit (MCNN-BiGRU) based on multi-source signals and applies it to the fault diagnosis and pattern recognition prediction of reciprocating compressor valves. Firstly, atom search optimization (ASO) is adopted to optimize the decomposition parameters of Swarm Decomposition (SWD) to obtain the ISWD algorithm, which is applied to decompose the multi-source signals of compressors to extract the oscillating components (OCs). Secondly, the correlation coefficient method is used to screen the OCs and conduct signal reconstruction, and various entropy feature values are extracted from the reconstructed signals to form an initial feature set. Then the t-SNE algorithm is employed to perform dimensionality reduction and fusion on the initial feature set, yielding a more concise and representative fused feature set. Finally, the fused feature set after dimensionality reduction and fusion is input into the MCNN-BiGRU model for training, so as to realize the pattern recognition and prediction of valve faults. The effectiveness and superiority of this method in the fault diagnosis of reciprocating compressor valves are verified through numerical simulation and experimental analysis.

Applied Sciences doi: 10.3390/app16115402

Authors: Kun Wang Yuanfeng Li Qiaoling Liu Kun Mao Yuan Yao

Mitigating heavy metal (HM) contamination in soil is vital for ecological and food security. Accurately mapping these pollutants and understanding their drivers are essential prerequisites for informed regional environmental governance. However, conventional spatial interpolation techniques used to estimate HM concentrations are susceptible to systematic biases and inadequate spatial resolution. To address these limitations, this study developed a novel hybrid model, termed HASM–ANN, coupling high-accuracy surface modeling (HASM) with artificial neural networks (ANNs). This approach generated high-resolution spatial distributions of HMs (As, Cd, Cu, Hg, Cr, and Pb) in agricultural soils of the Eastern Longquan Mountain region, Chengdu, China. Furthermore, the geographical detector (GD) and the Multiscale geographically weighted regression (MGWR) models were employed to explore driving mechanisms. Results indicate that HASM–ANN significantly outperformed conventional interpolations (ordinary/universal kriging, IDW) and HASM–coupled other machine learning downscaling methods. The proposed model demonstrated high predictive accuracy, yielding R2 values between 0.75 and 0.86, and consistently achieved a significantly lower RMSE across all targeted soil heavy metals compared to the HASM. Analysis of the explanatory power (q) revealed that soil As was primarily influenced by clay content (CC, q = 0.45) and available phosphorus (AP, q = 0.42), whereas Cd was mainly driven by AP (q = 0.51) and PM2.5 (q = 0.43). The spatial distribution of Hg was largely governed by soil organic matter (SOM, q = 0.53). Additionally, Cu concentrations were determined by SOM (q = 0.38), CC (q = 0.34), and pH (q = 0.31). Notably, Cr was significantly influenced by CC (q = 0.42), pH (q = 0.38), and elevation (q = 0.31), while Pb was further driven by SOM (q = 0.46) and PM2.5 (q = 0.39). By offering high-precision mapping and elucidating the underlying driving mechanisms, this research directly facilitates informed environmental governance to protect ecological integrity and public health.

Applied Sciences doi: 10.3390/app16115400

Authors: Chi-Tung Lai Sheau-Farn Max Liang

Scaffold-related falls remain a major safety concern in demolition projects, where temporary access systems are frequently erected, modified, used, and dismantled under changing structural and site conditions. These characteristics complicate risk prioritization because scaffold failures may involve interacting human, technical, organizational, and environmental factors. This study develops an expert-based risk prioritization framework for scaffold-related fall risks in demolition projects by integrating Failure Mode and Effects Analysis (FMEA), interval neutrosophic fuzzy (INF) theory, Decision-Making Trial and Evaluation Laboratory (DEMATEL), and Multi-Attributive Border Approximation Area Comparison (MABAC). Using the 4M1E perspective, namely Man, Machine, Material, Method, and Environment, 37 demolition-specific failure modes were identified through literature review and expert elicitation. Ten experts evaluated these failure modes using the SODE criteria, namely Severity, Occurrence, Detection difficulty, and Expected Cost impact. INF theory was used to represent uncertainty, hesitation, and judgmental variation in expert assessments. INF-DEMATEL was applied to examine interrelationships among the SODE criteria and derive interdependence-aware criterion weights, while INF-MABAC was used to rank the failure modes according to their distance from the Border Approximation Area. The framework was illustrated through an empirical application in Taiwan’s demolition industry. The results identified Severity as the most influential criterion. The highest-priority failure modes were insufficient safety awareness, improper scaffold-to-structure anchoring, and inadequate scaffold maintenance and inspection governance. Comparison with risk priority number (RPN)-based methods and sensitivity analyses using expert exclusion and Severity-weight variation showed that the ranking was generally consistent and reasonably stable under the tested conditions. The proposed framework provides a structured, uncertainty-aware decision-support procedure for identifying prevention priorities in demolition scaffold operations.

Applied Sciences doi: 10.3390/app16115399

Authors: Jinxiang Lai Deqing Wang Yifeng Zhao

With the wide application of artificial intelligence (AI) in the Internet of Vehicles (IoV), IoV is under pressure for data transmission and real-time sensing. Integrated sensing and communication (ISAC) is one of the key technologies to alleviate that pressure. Obstacles can cause communication disruptions and increased delays, hindering autonomous driving information acquisition and causing traffic hazards. The application of Reconfigurable Intelligent Surfaces (RISs) aims to solve this problem. This study focuses on RIS-assisted multi-base station (MBS) scenarios in the presence of obstacles. This study aims to maximize the communication rate, minimize the sensing error, and reduce the switching frequency by optimizing the RIS phase shift and beamforming. The problem is modeled as mixed integer nonlinear programming (MINLP) and further described as a Markov Decision Process (MDP). We use Long Short-Term Memory (LSTM) to predict the environmental state and propose two optimization algorithms, Multi-Factor Decision Deep Deterministic Policy Gradient (MFD-DDPG) and Mixed Discrete and Continuous Action DDPG (MDCA-DDPG). In the first algorithm, we consider multiple factors to make a switching decision and use DDPG to yield the optimal action. The second algorithm improves DDPG by outputting a discrete switching decision and a continuous optimized action simultaneously. Simulations show that the proposed algorithms significantly improve the system performance, and the communication rate is increased by more than 40% in specific multi-vehicle scenarios compared to the benchmark.

Applied Sciences doi: 10.3390/app16115398

Authors: Zhongbao Liu Wenjuan Zhao

Named entity recognition (NER) of ancient Chinese texts is the foundation for their development and utilization. Previous studies have focused on the data-driven methodology which tries to utilize the semantic features of ancient Chinese text. With the continuous accumulation of ancient Chinese linguistic resources and textual data, how to fully utilize the data resource and lexical knowledge related to ancient Chinese text with the help of new-generation information technology, so as to improve the ability of semantic comprehension and achieve good performance of NER, has become a great challenge to be solved. In view of this, this paper proposes a named entity recognition model for ancient Chinese text by fusing explicit feature and implicit feature (NERM), on the basis of extracting the explicit features and implicit features of ancient Chinese texts using a pre-trained model and a multi-head attention mechanism. In this model, the GuwenBERT model is introduced to extract the semantic features of ancient Chinese texts, namely the explicit features. The implicit features include relative positional relations, part-of-speech, and character radicals. The experimental results on the corpus GuNER 2023 show that the proposed model NERM achieves an F1 value of 90.67%, outperforming the existing models. The ablation experimental results show that implicit features provide a modest but meaningful improvement over explicit features, and implicit features can be arranged in order of importance as follows: character radicals, part-of-speech, and relative positional relations.

Applied Sciences doi: 10.3390/app16115397

Authors: Xianghui Xiao Pingyang Fan Yunchi Zhu Guangxu Jiang Chuankun Qiu Wenhao Zhao Min Wang

Pores, voids and cracks widely exist in rock materials, and the microstructural characteristics of rock are key to understanding its macroscopic behaviour because they strongly influence its mechanical behaviour in rock mass engineering. Owing to the importance of pores, voids and cracks to rock mechanical behaviour, studying the reconstruction of 2D microstructures can provide a basis for numerical simulation and engineering practice. The objective of this paper is to reconstruct the 2D microstructure, and a novel method for reconstructing the 2D microstructure is proposed. To reconstruct the 2D microstructure, an improved simulated annealing algorithm was used, and concurrent programming techniques were combined to fully utilize the computer resources. Moreover, to increase the convergence speed, a new technique that was efficient, accurate, and beneficial was proposed for reconstructing 2D microstructures. Through the simulation, the optimal number of neighbouring reverse pixels nreverse = 3 was determined and recommended. In this paper, a new approach for reconstructing 2D microstructures by using a raw sample is proposed. Moreover, 2D reconstruction was performed using both the improved simulated annealing (SA) algorithm and the conventional SA algorithm. The results show that the improved SA requires fewer iterations. This study provides a numerical basis for numerical simulation and engineering practice.

Applied Sciences doi: 10.3390/app16115396

Authors: Luigi Solazzi Nicola Zani Giorgio Donzella

Multi-layered shrink-fitted pressure vessels are critical for high-pressure applications, where structural integrity relies on inducing residual compressive stresses to mitigate operational tensile loads. This study presents a comprehensive analytical framework and automated sizing algorithms for both isotropic (metallic) and orthotropic (composite) thick-walled cylinders. Given fundamental design constraints, specifically the internal pressure, inner diameter and layer count, the models determine the optimal radial interferences required for assembly. For metallic configurations, geometric discretization is analytically derived from the Tresca yield criterion to guarantee uniform maximum equivalent stresses across all layers. For composite assemblies, a discrete optimization routine based on Classical Laminate Theory and the Tsai–Wu failure criterion is implemented to identify physically manufacturable repeated-sublaminate configurations, layer thicknesses and macroscopic equivalent properties. In both scenarios, interfacial contact pressures are derived by enforcing strict kinematic compatibility. The analytical stress fields and theoretical contact pressures are subsequently validated against Finite Element Method (FEM) simulations. Ultimately, the proposed algorithms provide an efficient and robust design tool capable of defining precise manufacturing tolerances and structural parameters for advanced high-pressure containment systems.

Applied Sciences doi: 10.3390/app16115395

Authors: Jiahao Pan George A. Kelley

The purpose of this study was to determine the association between the Timed Up and Go (TUG) test and mild cognitive impairment (MCI) and Alzheimer’s disease (AD). Cross-sectional studies were identified by searching five electronic databases and cross-referencing. Effect sizes were pooled using the inverse variance heterogeneity (IVhet) model, and certainty of evidence was evaluated using the Grading of Recommendations Assessment, Development and Evaluation (GRADE) instrument. Twenty-eight studies representing 1340 MCI, 1752 AD, and 37,561 healthy controls (HC) were included. Significantly greater completion time, in seconds, was observed for the MCI versus HC groups (X¯, 0.87, 95% CI, 0.38 to 1.37, p = 0.001; Q = 85.5, p < 0.001; I2 = 77.8%, 95% CI, 41.9 to 88.4%; 95% PI, −0.84 to 2.59) and AD versus HC groups when one major outlier was deleted from the model (X¯, 3.82, 95% CI, 2.57 to 5.07, p < 0.001; Q = 187.5, p < 0.001; I2 = 89.3%, 95% CI, 74.5 to 94.2.4%; 95% PI, −1.29 to 8.93). Based on GRADE, the overall certainty of evidence was considered very low. The current findings suggest very low-certainty evidence that the TUG test may be associated with MCI and AD when compared to HC. Additional, well-designed studies are needed before any level of conclusiveness can be established.

Applied Sciences doi: 10.3390/app16115394

Authors: Juan González-Quel Carlos Arnaiz del Pozo Ángel Jiménez Álvaro

Energy efficiency is a critical avenue for reducing carbonaceous emissions across fossil fuel value chains. Specifically, utilization of liquefied natural gas (LNG) exergy for power generation upon regasification in an import terminal offers the opportunity to partially retrieve the energy invested during liquefaction. Power generation arises as a promising avenue to accomplish this by using ambient air or seawater to supply heat to a working fluid, while the regasified LNG stream behaves as the heat sink of the thermal machine. However, a trade-off between cycle complexity (capital investment) and process efficiency exists. To identify it, in this work, three Rankine cycle configurations, which operate through indirect heat exchange without the need of fuel combustion, are analyzed with a consistent methodology from an exergo-economic perspective. Using a 2.13 mtpa LNG regasification terminal without LNG exergy utilization as the baseline for the techno-economic assessment, the simplest configuration consisting of a two-pressure level propane cycle (C3) achieved an exergy efficiency of 34.0% and a levelized cost of electricity (LCOE) of 89.4 €/MWh. A cycle carrying out an expansion of a portion of the regasified LNG and employing a CO2 loop for the high temperature range (C1CO2) achieved an exergy efficiency of 42.5% but with a higher LCOE of 99.7 €/MWh. Finally, the most capital-intensive design, comprising two stages with a hydrocarbon mixed refrigerant and propane as working fluids (MRC3), reached an efficiency of 55.2% and a cost of electricity of 118.5 €/MWh. The exergy analysis revealed that minimizing the MITA of cryogenic exchangers should be prioritized to improve cycle performance. However, even when large LNG regasification capacities (>6 mtpa) are considered, the most cost-effective solution (C3) generates profits during less than 45% of the time in the electricity market from 2024 of an LNG importing region such as Spain, indicating a relatively low economic potential for power generation without complementary heat sources.

Applied Sciences doi: 10.3390/app16115392

Authors: Stoil Kavalov Tanya Pehlivanova Miroslav Vasilev Zlatin Zlatev

Electromechanical systems operating under viscoelastic loads require precise modeling due to the highly nonlinear behavior of the load. An automatic bread machine is a practical example where dough represents a dynamic viscoelastic load sensitive to hydration. As found in this paper, increasing the water content leads to a decrease in the torque and the required mechanical power. An integrated approach combining MATLAB/Simulink and Simscape modeling, experimental measurements, and a PCA-based regression model is presented. The tests were conducted with three types of flour (type 500, type 1850, and rye–wheat) at hydrations of 52%, 58%, and 63% with over 6000 measurements recorded for each combination. The regression models achieve moderate predictability (R2 = 0.64–0.96) model performance that varies across flour types. Increasing the dough hydration from 52% to 63% reduces the torque by approximately 22–46% across the tested flour types, while the angular velocity rises slightly (from about 147.9 to 151.9 rad/s). A descriptive decrease in energy consumption of up to around 6% was observed within the sampled batches with the system efficiency remaining within a narrow range around η ≈ 0.67. Within the studied levels (52–63%), the minimum load was observed at 58%. The proposed integrated model reliably describes the interaction between the electric motor, the mechanical gear, and the viscoelastic load, and it offers a basis for energy optimization and the implementation of low-cost sensor systems for intelligent control in the bread-making process.

Applied Sciences doi: 10.3390/app16115393

Authors: Ning Tian Meng Li Yang Liu Haonan Zhang Xiaozhou Zhou

Discontinuity-controlled rock masses in underground caverns are prone to block formation and instabilities during construction, motivating rapid tools for stability feedback analysis. Fast screening of discontinuity-controlled block hazards in underground caverns is addressed through a physics-consistent machine learning framework for identifying block formation and assessing instability conditional on formation. Cavern geometry, rock mass properties, and multiple joint sets are parametrically encoded, and physically and geometrically consistent data augmentation is performed. Three-dimensional discrete element batch simulations provide automated labels for block formation (yB) and instability (yU) (FoS < 1), forming a training dataset. Twenty-three raw features with a 66-dimensional engineered feature set are compared and multiple classifiers using PR/ROC curves, confusion matrices, and a BAcc-based threshold strategy are evaluated. Compared with raw inputs, the engineered features generally improve AUC-based ranking and balanced discrimination, especially for the RF model, although threshold-dependent recall and F1 trade-offs are observed for some learners. Random forests show consistently robust results. Benchmarking against an independent engineering reference based on stereographic projection and limit-equilibrium analysis gives 91% agreement for block-formation identification and 85.2% agreement for conditional instability identification. The trained model is integrated into an engineering platform to support batch screening of three- and four-plane combinations with 3D visualization outputs.

Applied Sciences doi: 10.3390/app16115390

Authors: Chunju Zhao Zhiyu Liu Fang Wang Yihong Zhou Jun He Huawei Zhou Zhipeng Liang Lei Lei

In the construction of conventional concrete high arch dams in high-altitude regions with large temperature variations, the prolonged and cold winters often force the suspension of concrete pouring, severely constraining the overall schedule. To address this limitation, this paper breaks away from the conventional winter-shutdown scheme by proposing a new technique: continuous construction under low-temperature conditions. It can adapt to large temperature variations, and this study develops a corresponding construction schedule simulation model for quantitative evaluation and scheme optimization. First, the influence of large diurnal temperature variations on high-altitude concrete pouring was analyzed. Based on this, a dynamic pouring technique for sub-blocks is proposed—thin-layer pouring during positive temperatures and insulation curing during negative temperatures—with the aim of transforming discrete climatic windows into a continuous construction period. Second, to accurately simulate this complex spatial partitioning and temporal scheduling process, a customized schedule simulation model based on discrete-event simulation (DES) theory was developed. The model incorporated meteorological recognition at low temperatures, dynamic dam-block partitioning, and sub-block pouring scheduling. Finally, a high arch dam on a plateau in Southwest China was used as an engineering case to compare two construction schemes: the low-temperature shutdown scheme and the continuous construction scheme. After validating the simulation model under parameter assumptions such as ideal resource availability and stable annual climate patterns, the results showed that the continuous construction scheme achieves a monthly average pouring volume of 33,721 m3 during the period with large diurnal temperature variations, which accounts for 42.48% of the average monthly pouring volume during the normal construction period. Compared to the low-temperature shutdown scheme, the coefficient of variation of the monthly pouring intensity decreases by about 40%, and the total construction period is shortened by approximately ten months. This demonstrates the potential for schedule optimization for continuous winter construction in simulation.

Applied Sciences doi: 10.3390/app16115391

Authors: Qiuying Guo Zhengyu Wang Dewei Li Yingjun Sun Jian Zhang Heng Liu

Precipitable water vapor (PWV) is a pivotal parameter in the measurement of atmospheric water vapor content. Interferometric Synthetic Aperture Radar (InSAR) is capable of retrieving PWV with high accuracy and spatial resolution. However, the limitations imposed by factors such as low coherence and phase distortion prevent the monitoring of PWV over the sea surface by InSAR. For this problem, a joint InSAR/Fengyun sea–land cooperative PWV construction method is proposed using constraints from the Global Navigation Satellite System (GNSS) and ERA5 reanalysis data. The Jiaodong Peninsula of China was selected as the research area. The PWV over the land of the Jiaodong Peninsula was obtained by GNSS/ERA5/InSAR. The PWV over the nearby sea was obtained by Fengyun-4A (FY-4A). The PWV reconstruction over the sea and land region is realized by means of unified reference correction and transition processing. The results indicate that InSAR PWV and FY-4A PWV show good agreement with GNSS PWV, with R2, MAE and RMSE values of 0.955, 1.86 mm and 2.32 mm for InSAR PWV, and 0.961, 1.90 mm and 2.28 mm for FY-4A PWV, respectively. The fused PWV significantly improves spatial completeness while preserving fine spatial structures, with an annual PWV range of 0.35–59.74 mm and a clear seasonal cycle from 4.16 mm in winter to 33.24 mm in summer. The results effectively capture the coastal moisture transition, with the strongest PWV gradient reaching 0.823 mm/km in the 0–10 km coastal zone during the warm season. PWV also shows the strongest correlation with dew point, with a Spearman correlation coefficient of 0.94. This study overcomes the limitation that PWV is restricted to land areas and provides reliable data support for the analysis of water vapor structures in coastal regions.

Applied Sciences doi: 10.3390/app16115389

Authors: Xuehua Song Li Han Yi Zhu Qianxiang Wei Zijun Yang Xiaoming Jiang

Agricultural AI agents play a crucial role in the evolution of smart agriculture, from single-point automated applications to intelligent systems driven by tasks, collaborative decision-making, and closed-loop execution. However, their practical implementation still faces key challenges, such as heterogeneous agricultural data processing, insufficient cross-scenario generalization ability, complexity of multi-agent collaboration, difficulties in integrating software and hardware, and insufficient security and trust guarantees in real agricultural environments. This paper presents a systematic review of the architecture design, business processes, key technologies, and future challenges of agricultural AI agents. Agricultural AI agents are classified into two types: virtual agricultural AI agents and embodied agricultural AI agents. The paper summarizes a four-layer system architecture consisting of the infrastructure layer, agent management layer, agent collaboration layer, and application layer. The paper also analyzes the model capabilities required by agricultural AI agents from four typical business dimensions: perception and state understanding, knowledge memory and experience management, reasoning decision-making and task planning, and collaborative execution and resource scheduling. This research shows that technologies such as multimodal perception, knowledge graphs, retrieval-enhanced generation, digital twins, reinforcement learning, and multi-agent collaboration can provide important support for agricultural AI agents to enhance their environmental understanding, knowledge reuse, autonomous decision-making, and physical execution capabilities. Future research should focus on robust perception in open environments, long-term memory and knowledge evolution, reliable multi-agent collaboration, edge-cloud collaborative deployment, and secure and trustworthy human–machine collaboration. Integrating agricultural domain knowledge with intelligent agent technology is an important direction for promoting the large-scale, adaptive, and sustainable application of agricultural AI agents.

Applied Sciences doi: 10.3390/app16115387

Authors: Joseph B. Bernstein

Reliability degradation in semiconductor devices originates from microscopic stochastic processes such as defect motion, diffusion, bond rearrangement, and charge trapping occurring under electrical and thermal stress. Experimental degradation measurements, however, often exhibit smooth empirical scaling behavior, particularly power-law time dependences extending across many orders of magnitude in time. This tutorial reviews the thermodynamic and kinetic foundations underlying these observations and explains how empirical power-law degradation behavior can emerge from the collective interaction of many microscopic stochastic processes. The discussion begins with irreversible thermodynamics, random walk transport, diffusion, and Arrhenius kinetics and then connects these microscopic concepts to the macroscopic degradation trends commonly observed in semiconductor reliability experiments. Attention is given to the interpretation of stress-dependent power-law degradation kinetics and their implications for accelerated lifetime extrapolation. Practical limitations associated with conventional logarithmic degradation analysis are examined, including baseline sensitivity, logarithmic instability near the measurement floor, and systematic curvature that may remain hidden despite high goodness-of-fit metrics. Methods based on transformed-coordinate linearization and curvature-sensitive extraction are discussed together with their implications for time-to-failure extrapolation and activation-energy interpretation. Experimental studies of phenomena such as bias temperature instability frequently show degradation behavior in which the time exponent depends systematically on voltage and temperature stress conditions. Under such conditions, the reciprocal exponent m=1/n can significantly amplify stress acceleration during lifetime extrapolation. This work provides a conceptual framework connecting microscopic stochastic degradation physics with the empirical methods commonly used in practical semiconductor reliability analysis and long-term lifetime prediction.

Applied Sciences doi: 10.3390/app16115388

Authors: Dahai Guo Guangchen Liu Jianwei Zhang Guizhen Tian Sufang Wen Zicheng He Yan Wang

To address DC-link voltage fluctuation, active-power imbalance between the machine side and the grid side, and double-frequency distortion in the grid current of a flywheel energy storage system (FESS) under symmetrical and asymmetrical voltage sag faults, this paper proposes a coordinated control strategy for the machine-side and grid-side converters to enhance low-voltage ride-through (LVRT) capability. Taking the DC-side energy imbalance as the coordination criterion, the machine-side converter adopts an online active-current-command reconstruction method based on cascaded limiting of DC-link voltage deviation. Under reactive-power-priority support and constrained active-power output on the grid side, the FESS can actively adjust its active-current command according to the DC-side energy state, thereby suppressing DC-link overvoltage/undervoltage and restoring the power balance between the machine side and the grid side. On the grid side, an improved linear active disturbance rejection control (LADRC) is introduced into the current inner loop. By optimizing the structure of the extended state observer, the observation and compensation capability for double-frequency disturbances is enhanced, thus improving grid-current quality under asymmetrical faults. In this way, power rebalancing between the machine side and the grid side, DC-link voltage stabilization, and grid-current disturbance suppression are incorporated into a unified coordinated control framework. Hardware-in-the-loop experimental results show that the proposed strategy can maintain DC-link voltage stability during both symmetrical and asymmetrical voltage sags, while keeping the maximum grid-current total harmonic distortion (THD) below 0.13%. Under asymmetrical voltage sag, the improved LADRC reduces the maximum interphase peak-current deviation from approximately 52 A under conventional PI control to 4.57 A, corresponding to a reduction of about 91.2%. These results indicate that the proposed strategy can effectively enhance DC-link voltage stabilization and improve grid-current quality during faults.

Applied Sciences doi: 10.3390/app16115386

Authors: Seung-han You Jin-kook Kim Sung-wook Kim Jae-heum Moon Won-woo Kim

Structural damage under blast conditions is significantly influenced by high-velocity primary fragments, which can induce severe local penetration and threaten the safety of protective structures. In this study, the penetration behavior of concrete subjected to primary fragment impact was investigated using the Adaptive Smoothed Particle Hydrodynamics (SPH) method implemented in LS-DYNA. The numerical model was first validated against gas-gun experimental results, demonstrating improved prediction accuracy compared to the conventional Lagrangian approach. In particular, the Adaptive SPH method effectively mitigated premature element erosion, resulting in a residual velocity prediction error of approximately 5.9%. Based on the validated model, a parametric study was conducted to evaluate the effects of fragment mass (3.7–44 g) and impact velocity (750–1000 m/s) on concrete penetration behavior. The results showed that most fragments were capable of perforating a 100 mm thick concrete wall at velocities above 1000 m/s, while partial penetration occurred only under limited conditions. Furthermore, a strong logarithmic relationship between fragment mass and residual velocity was identified, and predictive equations with high reliability (R2 ≥ 0.98) were proposed. These findings demonstrate the applicability of the Adaptive SPH approach for realistic penetration analysis and provide practical insights for the design and safety assessment of protective concrete structures.

Applied Sciences doi: 10.3390/app16115385

Authors: Catarina M. Amaro Maria António Castro Ana M. Amaro

Understanding muscle activation patterns during sport-specific skills is essential for optimizing performance and training strategies. In basketball, upper limb actions such as passing and receiving require precise coordination and effective neuromuscular control. The main goal of this study was to analyze and compare the muscle activity of the biceps brachii and triceps brachii during the execution and reception of the two-handed chest pass in basketball players with different levels of competitive experience. Surface electromyography (EMG) data were collected from 14 federated athletes, aged between 11 and 29 years, using the BioSignal Plux system. Participants were allocated into two groups according to their playing experience. Muscle activation was analysed in terms of activation time (AT) and percentage of muscle activation (%MA), normalised to maximum voluntary contraction (MVC). A linear mixed model was used to evaluate the effects of experience level, limb dominance, and their interaction while accounting for repeated measures within participants. No significant differences were observed between dominant and non-dominant limbs for any variable. Significant differences between experience/age groups were identified mainly in the triceps brachii, particularly for activation time in the lateral head and %MA in the long head. In general, more experienced/aged athletes demonstrated higher levels of neuromuscular activation and shorter activation times, suggesting different motor control strategies. A significant positive association was found between years of practice and %MA of the long head of the triceps brachii. These findings provide novel insights into neuromuscular recruitment during both the execution and reception phases of the basketball chest pass and may inform training strategies aimed at enhancing technical efficiency across developmental stages.

Applied Sciences doi: 10.3390/app16115384

Authors: Marco Antonio Villagómez-Cuéllar Elisabeth Bilbao-García Unai Duoandicoechea Natalia Villota

The removal of persistent pharmaceutical compounds such as carbamazepine (CBZ) by advanced oxidation processes (AOPs) remains a major challenge in water treatment, particularly in relation to understanding the operating conditions governing reaction kinetics and transformation pathways. In this context, this study aims to evaluate the effect of ozone concentration on the kinetics and mechanistic regimes of CBZ ozonation in aqueous solutions. Ozonation experiments were conducted in an aqueous solution at an initial CBZ concentration of 50.0 mg/L, using inlet ozone concentrations between 1.9 and 58.5 g/m3 under controlled conditions. CBZ degradation followed apparent pseudo-first-order kinetics under the studied conditions, with the corresponding apparent rate constant increasing linearly with the inlet ozone concentration. At ozone concentrations ≥ 15.7 g/m3, rapid CBZ removal was observed, together with high dissolved ozone levels, accelerated loss of aromaticity, and transient formation of colored oxidation intermediates, which were subsequently degraded. In contrast, low ozone concentrations led to ozone-limited kinetics and slower aromatic breakdown. The pH evolution revealed two distinct kinetic regimes, transitioning from oxidant-limited to reaction-controlled behaviour and stabilizing at pH 4.3. These findings may provide guidelines for optimizing ozone-based treatment processes. The insights gained may be applied to the design, scale-up, and operation of advanced and hybrid oxidation systems.

Applied Sciences doi: 10.3390/app16115383

Authors: Bilge Atay Karlıdağ Tülün Malkoç Seval Eminoğlu Küçüktepe

Background: Artificial intelligence (AI)-supported learning environments are increasingly used in music education; however, evidence regarding their effectiveness in solfège instruction remains limited. This action research study evaluated the effects of AI-supported solfège instruction on undergraduate students’ attitudes and performance. Materials and Methods: This action research study included 36 undergraduate students enrolled in a conservatory program. A 10-week AI-supported solfège training was implemented using the EarMaster intelligent tutoring system. Data were collected through a solfège attitude scale and a performance test administered before and after the intervention, along with a delayed retention test. Results: Following the intervention, significant improvements were observed in both attitude scores (104.6 vs. 117.3, p < 0.001) and performance scores (32.8 vs. 51.52, p < 0.001). However, retention test scores showed a significant decline after a no-practice period (51.5 vs. 48.8, p < 0.001). Qualitative findings indicated increased motivation, engagement, and individualized learning opportunities. Conclusions: AI-supported solfège instruction improved students’ performance and learning attitudes in higher music education. The findings suggest that adaptive feedback, individualized practice, and continuous engagement may contribute positively to auditory skill development and student motivation. However, sustained practice remains necessary for long-term retention. Artificial intelligence-supported systems should therefore be integrated as complementary tools alongside teacher-guided instruction to support more flexible and personalized learning environments.

Applied Sciences doi: 10.3390/app16115382

Authors: Sophia Mosalla So Yeon Jeon Jeong-Dong Kim Taek Lee Jung-Been Lee

Caregiver burden is a critical public health issue among individuals providing long-term care for family members with cognitive impairment and is closely associated with physiological stress. However, conventional assessments primarily rely on self-reported measures, which provide limited insight into underlying physiological processes and temporal dynamics. To address this limitation, we investigate the use of wearable-derived physiological and behavioral signals for the objective assessment of caregiver burden. We analyzed multimodal data collected from 78 informal caregivers using Fitbit devices, including heart rate, activity patterns, sleep architecture, and variability-based features aggregated over a 7-day window. Associations with caregiver burden, measured by the Zarit Burden Interview (ZBI), were evaluated using Spearman correlation with false discovery rate correction and multivariable regression models adjusted for demographic and clinical covariates. Variability in REM sleep duration showed an exploratory negative association with caregiver burden (ρ = −0.32, p = 0.004), with similar but weaker nominal patterns observed for light and total sleep variability. In multivariable analysis, maximum daily heart rate was positively associated with caregiver burden (β = 2.59, p = 0.005), although overall model performance was modest (R2 = 0.32, adjusted R2 = 0.08) and the full model did not reach statistical significance. These findings demonstrate that multimodal wearable signals may provide complementary objective insights into caregiver burden and support the evaluation of sleep variability and heart rate measures as candidate indicators.

Applied Sciences doi: 10.3390/app16115381

Authors: Hailong Li Yun Wang Jian Liu Haiyang Liu

As the area occupied by static random access memory (SRAM) continues to increase in advanced integrated circuits, SRAM yield has become a critical factor that directly constrains chip cost and manufacturing efficiency. Conventional SRAM redundancy configuration methods are largely based on ideal random-defect assumptions and therefore cannot accurately characterize the systematic defects that are widely observed in advanced technology nodes. This mismatch often leads to suboptimal redundancy allocation with respect to the actual failure distribution. To address this issue, this paper proposes a data-driven SRAM redundancy optimization method for mass-production applications. The proposed method integrates defect distribution modeling, systematic defect identification, test-algorithm signature extraction, and physical failure analysis (PFA) into a closed-loop framework of test, diagnosis, localization, and optimization. Experimental results based on 7 nm mass-production chips demonstrate that the proposed method can effectively identify systematic defects, achieving an initial PFA localization hit rate close to 100% for single stuck-at faults while significantly improving failure analysis efficiency. Further redundancy evaluation shows that, after the major systematic defects are removed, the required redundancy can be reduced from two-row/two-column redundancy to only single-column redundancy while still covering all repairable failures, thereby improving both area efficiency and manufacturing economy. The proposed method provides a practical engineering solution for SRAM redundancy planning, process tuning, and yield improvement in advanced technology nodes.

Applied Sciences doi: 10.3390/app16115380

Authors: Ionel Cristian Vladu Nicu George Bîzdoacă Ionica Pirici Tudor-Adrian Bălșeanu Eduard Nicușor Bondoc

Contemporary neuroscience has generated extensive empirical insights into perception, memory, prediction, valuation, and consciousness. However, it still lacks an explicit operational architecture capable of explaining how these processes emerge from a unified computational mechanism. This work introduces DIME (Detect–Integrate–Mark–Execute), a unified operational architecture in which perception, memory, valuation, and conscious access are treated as components of a single recurrent computational cycle. The framework is organized around four core elements: engrams, defined as distributed recurrent neural structures that support multiple activation trajectories rather than static memory traces; execution threads, representing temporally extended, causally coherent trajectories of neural activity; marker systems, corresponding to neuromodulatory and limbic mechanisms that regulate value, selection, plasticity, and trajectory competition; and hyperengrams, large-scale integrative states associated with global coordination and conscious access. Within this formulation, DIME provides a mapping between local neural assemblies, temporal sequence dynamics, value-based modulation, and large-scale network integration. Rather than treating perception, memory, and decision-making as partially independent processes, the framework interprets them as different expressions of a single operational loop acting across multiple spatial and temporal scales. The proposed architecture is consistent with empirical findings on hippocampal indexing, recurrent cortical processing, neuromodulatory control, and large-scale network dynamics, while remaining sufficiently general to support applications in artificial intelligence and robotics. Unlike frameworks centered on prediction, memory storage, or global broadcasting, DIME proposes that cognition arises from the recurrent interaction between executable representational structures, trajectory-based processing, value-guided selection, and dynamic large-scale integration. The framework generates explicit and falsifiable predictions regarding context-dependent neural trajectories, marker-mediated state transitions, and large-scale network reconfiguration. In this sense, DIME is not intended as a metaphorical synthesis, but as a testable architectural hypothesis for neuroscience and biologically inspired cognitive systems. Beyond theoretical neuroscience, the framework is also positioned as a transferable design-level reference model for adaptive AI systems, autonomous robotics, and cognitively informed engineering architectures operating in dynamic environments.

Applied Sciences doi: 10.3390/app16115379

Authors: Duanne Engelbrecht Karim Djouani Nico Steyn Gustave Udahemuka

The Performance Efficiency Factor (PEF) is a key composite metric in broiler production that integrates livability, average body weight, feed conversion ratio (FCR) and age. While tree-based machine learning models have shown promising results for live-weight prediction, they often struggle with temporal dependencies and sparse mortality data. This study evaluates five tree-based algorithms, Decision Tree, Random Forest, Extreme Gradient Boosting (XGBoost), Light Gradient Boosting Machine (LightGBM) and Category Boosting (CatBoost) on daily sensor data from commercial broiler farms. Data were stratified into six age categories, pre-processed with outlier retention and feature engineering, and assessed using R2, RMSE, MAE, paired t-tests for statistical significance and SHapley Additive exPlanations (SHAP) for explainability. Cross-farm validation was performed to assess generalizability. The models achieved high accuracy for live-weight prediction (R2 up to 0.97, RMSE as low as 9.23 g in mid-cycle categories), with Random Forest and LightGBM performing best. Mortality prediction remained challenging (R2 −1.63 to 0.53) due to its sparse and stochastic nature. Nevertheless, Day 21 to Day 28 PEF forecasts showed relative errors of only 4.42–5.40%, as live-weight predictions dominated the PEF calculation. SHAP analysis consistently identified bird age, feed intake per bird and temperature as the most influential predictors. Tree-based models offer a robust, interpretable and computationally efficient solution for live-weight and PEF forecasting in commercial broiler production. The findings support proactive farm management and highlight the need for hybrid approaches to improve mortality prediction.

Applied Sciences doi: 10.3390/app16115378

Authors: Monika Michalak Aneta Ostróżka-Cieślik Magdalena Hartman-Petrycka Anna Stolecka-Warzecha Sławomir Wilczyński

Near-infrared radiation contributes to photoaging through oxidative stress and matrix metalloproteinase activation. Botanical extracts with antioxidant properties may offer additional protection beyond conventional UV filters. To evaluate the effect of hydrogel formulations containing Rosmarinus officinalis and Crataegus monogyna extracts on the directional reflectance of human skin across various spectral ranges. Directional reflectance was measured on the forearm skin of healthy female volunteers before and after application of a base hydrogel and hydrogels containing plant extracts. Hyperspectral imaging was used across spectral ranges of 335–2500 nm. To assess the application properties, rheological and textural evaluation of extract-based hydrogels was performed. The obtained results are satisfactory and indicate the expected application effectiveness of hydrogels with C. monogyna and R. officinalis extracts. Significant reductions in skin reflectance were observed in the IR spectrum after application of both botanical formulations. Median reflectance decreased by 3.5% with rosemary and 2.3% with hawthorn in the 1000–1700 nm range, and by 17.8% and 20.3% respectively in the 1700–2500 nm range. No statistically significant changes were observed in the UV or visible light ranges. Hydrogels enriched with R. officinalis and C. monogyna extracts reduced infrared reflectance of the skin, suggesting potential as adjunctive agents in photoprotection. These findings support further investigation into extract-based formulations for IR-related skin damage prevention.

Applied Sciences doi: 10.3390/app16115377

Authors: Anna Skic Zbigniew Stropek Kamil Drabik Pavol Findura Miroslav Prístavka

Bone tissue exhibits viscoelastic behavior, and stress relaxation analysis provides valuable insight into its time-dependent mechanical performance. This study aimed to evaluate the effects of allicin supplementation on the mechanical and viscoelastic properties of tibial bone in Bovans Brown laying hens. Tibiae of 68-week-old hens were subjected to three-point bending tests to determine quasi-static mechanical properties, as well as stress relaxation tests conducted at three deformation velocities (0.1, 1, and 10 mm/s). Stress relaxation behavior was described using a five-parameter generalized Maxwell model. Allicin supplementation resulted in significantly greater bone stiffness and work to fracture compared with the control group. Stress relaxation tests demonstrated a rate-dependent viscoelastic response in all samples. Bones from allicin-supplemented hens exhibited higher force levels during the relaxation period and reduced stress decay, indicating enhanced elastic behavior and decreased viscous deformation. Model analysis revealed a tendency toward higher equilibrium modulus values and longer relaxation times in the supplemented group, particularly at low deformation velocities. However, no statistically significant differences were observed for E0, τ1, and τ2 between the tested groups. These findings suggest that allicin supplementation may improve bone mechanical and viscoelastic properties in laying hens.

Applied Sciences doi: 10.3390/app16115376

Authors: Bartłomiej Zieniuk Katarzyna Zielińska Marta Siol Diana Mańko-Jurkowska Andrzej Bryś Joanna Bryś

Oat beverages are widely consumed as dairy alternatives; however, there is still limited understanding of how the source and structural organization of added lipids influence their physicochemical functionality and sensory perception. This study aimed to evaluate whether the type of lipid source used in commercial oat-based products can explain variability in lipid quality and consumer acceptability. Lipid fractions were extracted from 10 commercial oat-based products available on the Polish market, including 5 ready-to-drink beverages and 5 powdered products reconstituted with water. The extracted lipids were analyzed for fatty acid (FA) composition, positional distribution of FAs in triacylglycerols, melting behavior, and oxidative stability, complemented by sensory evaluation. Marked differences were observed among samples, primarily driven by the lipid source rather than product form. Products containing sunflower oil exhibited a favorable FA profile characterized by a high proportion of unsaturated FAs and relatively good oxidative stability. In contrast, oat-only formulations showed lower oxidative stability and reduced sensory performance, particularly in terms of taste and texture. The sample containing coconut fat demonstrated the highest oxidative stability (τmax = 333.48 min) but the least favorable nutritional profile due to a predominance of saturated FAs (85.43% SFA). The highest overall sensory acceptance was recorded for sample L1 (overall desirability = 7.00). Overall, the findings demonstrate that lipid source is a key determinant of the nutritional, physicochemical, and sensory properties of oat beverages, while product format (liquid vs. powder) plays a secondary role. These results address the knowledge gap regarding the relationship between lipid origin and functional performance in plant-based beverages and highlight formulation strategy as a critical factor in product optimization.

Applied Sciences doi: 10.3390/app16115375

Authors: Yuruo Wang Xiangchao Shi

In gas drilling, the local annular contraction caused by a drillpipe tool joint can markedly reduce cuttings’ carrying capacity and increase the risk of localized blockage and sand bridging near the tool-joint region, thereby threatening hole cleaning and drilling safety. To investigate this problem, a three-dimensional CFD–DEM two-way coupling model was established by considering the geometric features of the drillpipe tool joint and gas–solid interaction. The effects of gas mass flow rate, solids feed rate, and particle diameter on local cuttings’ transport states and annular pressure-drop responses near the tool joint were systematically analyzed. The results show that three typical local transport states can develop near the tool-joint region, namely continuous passage, fallback, and clogging accompanied by sand-bridge formation. Fallback cases occur only within a finite interval around the critical gas mass flow rate for cuttings’ transport. Under the geometric and operating conditions considered in this study, localized clogging first appears when the particle diameter reaches approximately 10.5 mm, and the proportion of clogging cases increases rapidly with a further increase in particle diameter. Increasing the solids feed rate intensifies particle retention, accumulation, and collision near the tool joint, promotes earlier clogging, and markedly narrows the operating range of continuous passage; stable clogging is difficult to form when the solids feed rate is below 8 kg/s. Distinct annular pressure-drop histories correspond to different local transport states, with low amplitude fluctuation for continuous passage, repeated pulsation for fallback, and sustained growth in pressure difference magnitude for developing clogging accompanied by sand bridge formation. These results demonstrate a clear correspondence between local transport states near the tool joint and annular pressure-drop responses under the investigated geometry and operating window. They provide a mechanism-level basis for interpreting localized blockage near the drillpipe tool joint, while quantitative field application requires calibration for the specific annular clearance, monitoring interval, gas-injection condition, and cuttings’ loading condition.

Applied Sciences doi: 10.3390/app16115374

Authors: Manuel J. C. S. Reis

The increasing penetration of converter-interfaced renewable energy resources is accelerating the transition of conventional distribution networks toward hybrid AC/DC architectures, where photovoltaic generation, battery energy storage, electric mobility, and mixed AC/DC loads are coupled through multiple power electronic interfaces. While these architectures offer important advantages in flexibility and integration efficiency, they also introduce tighter interactions between AC-side and DC-side operating behavior, making coordinated assessment increasingly important under variable operating conditions. Despite growing interest in hybrid AC/DC systems, comparative studies that jointly examine system-level stability and voltage-quality-related behavior across renewable penetration levels and stressed operating scenarios remain limited. This paper proposes a reduced-order comparative screening framework for renewable-rich hybrid AC/DC distribution systems, using stability- and voltage-quality-related indicators based on a representative reduced-order benchmark model. The adopted framework combines scenario-based simulation with unified AC-side, DC-side, transient, and composite performance indicators to evaluate how different converter coordination strategies influence operating robustness under renewable intermittency, abrupt load changes, converter operating-point variations, and different renewable penetration levels. The considered indicators include voltage deviation, overshoot, violation duration, transient fluctuation, converter utilization, and composite operating-robustness measures; they are intended as system-level voltage-dynamics proxies rather than as a complete harmonic or standards-based power-quality assessment. The results indicate that adaptive coordinated control provides the strongest DC-side robustness under stressed conditions, whereas droop-based coordination often offers a favorable practical compromise between AC-side and DC-side performance. The analysis also reveals a clear trade-off between DC-side regulation and AC-side voltage-quality-related behavior, highlighting the need for joint multi-domain evaluation. In particular, the improved DC-side robustness obtained with adaptive coordination is accompanied by slightly higher AC-side voltage-quality-related deviations in several scenarios. Within the scope of the adopted reduced-order benchmark, the proposed framework provides a practical and reproducible basis for identifying critical operating regions and for supporting higher-fidelity future studies on robust renewable integration in hybrid AC/DC distribution networks.

Applied Sciences doi: 10.3390/app16115373

Authors: Hongwei Deng Jingbo Xu Jun Shen Zeru Cui Junren Deng

As large-scale continuous geological weak planes widely developed in deep mines, through-going discontinuities have a significant impact on the distribution of deep rock mass stress fields and the deformation and failure modes of surrounding rock and are the core factors restricting the safe and efficient extraction of deep resources. In this paper, indoor tests and numerical analysis methods were adopted to study the mechanical strength evolution characteristics and displacement changes of monitoring points of rock mass with prefabricated through-going discontinuities under horizontal and vertical pressure conditions, and the influence laws of different structural thicknesses and distances on the stress characteristics of engineering cavities were obtained. The results show that the introduction of through-going discontinuities can significantly reduce the peak strength and elastic modulus of rock specimens and change the circumferential strain. The thickness, distance, and relative position of through-going discontinuities jointly determine the strength degradation characteristics of the specimens. The larger the structural thickness (3 mm, 5 mm, and 10 mm), the weaker the strength degradation effect; when the through-going discontinuities are located above the cavity, the degradation effect is more significant. The relative position between the structure and the cavity determines the overall level of the stress field, while the structural thickness and distance (20 mm, 25 mm, and 30 mm) regulate the stress distribution and transfer path. When the through-going discontinuities are located above the cavity, the vertical stress of the surrounding rock decreases significantly with the increase of thickness, the horizontal stress remains basically stable, and the overall presents a high stress concentration characteristic. When the through-going discontinuities are located below the cavity, the vertical stress of the surrounding rock is generally low and changes gently; the horizontal stress increases gradually with the increase of thickness, and the overall stress level is much lower than that of the upper working condition.

Applied Sciences doi: 10.3390/app16115372

Authors: Harun Güçlü

In battery electric vehicles (BEVs), range-extended electric vehicles (REEVs) are gaining prominence due to range limitations, long charging times, and limited charging infrastructure. Range losses are particularly evident under extreme climate conditions, necessitating the development of efficient range-extender (RE) systems. In this study, a liquefied petroleum gas (LPG)-fueled, recuperator-equipped Micro Gas Turbine (MGT) was modeled as a standalone range-extending power unit using the Simcenter simulation environment, and its thermodynamic performance was examined under extreme climate conditions. While existing MGT studies in the literature generally focus on diesel-fueled systems, this study fills a significant gap in the literature by modeling the effects of using low-carbon, high-energy-density LPG. The performance of the MGT system was analyzed in extreme cold (−10 °C), standard (20 °C), and hot (45 °C) climates; at three different turbine inlet temperatures (1000, 1100, and 1250 K); and at three recuperator effectiveness settings (0.75, 0.85, and 0.95). The developed MGT system achieved a maximum thermal efficiency of 41.1% and a specific fuel consumption (SFC) of 188.67 g/kWh under cold climate conditions of −10 °C (263.15 K), a turbine inlet temperature (TIT) of 1250 K, and a recuperator effectiveness of 0.95. Consequently, specific CO2 emissions were reduced to 566.01 g/kWh. The study’s most significant contribution to the literature is that the developed system offers high thermal efficiency, low fuel consumption, and low emissions under extremely cold climate conditions (−10 °C), where electric vehicle batteries typically experience performance and range loss. The LPG-fueled micro gas turbine with a recuperator demonstrates the potential to serve as an efficient, low-emission and competitive auxiliary power unit (APU) for range-extender applications, particularly under extreme climatic conditions.

Applied Sciences doi: 10.3390/app16115370

Authors: Xiao Xuan Qi Ping Bobo Zhang

Dust and coal mine gas in deep mines are highly prone to causing fires, and the cyclic high temperatures generated by such fires are one of the key factors contributing to the instability of deep rock structures. To research the dynamic splitting tensile mechanical properties of sandstone subjected to high-temperature cycling, impact splitting tensile tests were performed on sandstone specimens under normal temperature and after high-temperature cycling treatments ranging from 250 °C to 900 °C using a split Hopkinson pressure bar (SHPB) with increasing cyclic temperature. The average dynamic tensile strength of sandstone specimens declines following a quadratic function, dropping from 18.07 MPa at T = 150 °C to a minimum value of 3.08 MPa, representing a maximum reduction of 82.96%. The dynamic strain and average strain rate exhibit increasing trends following exponential and logarithmic functions, respectively, while the dynamic elastic modulus exhibits a logarithmic declining trend. As the cyclic temperature grows, the degree of fragmentation of the specimens intensifies, transitioning from axial splitting failure to pulverization failure, with fragment size decreasing and fractal dimension exhibiting increasing trends. For temperatures between 450 °C and 600 °C, the dynamic tensile strength, dynamic strain, average strain rate, dynamic elastic modulus, average particle size, and fractal dimension all show a distinct interval behavior. As the cyclic temperature rises, the incident, reflected, and transmitted energies gradually decline. A higher fragmentation energy density corresponds to more severe specimen fragmentation, and the average fragment size follows a negative quadratic relationship with fragmentation energy density, which effectively quantifies the dynamic splitting tensile fragmentation behavior of rock. The findings of this study regarding the dynamic behavior and damage evolution of sandstone under cyclic high-temperature conditions can serve as a reference for assessing rock mass stability in high-temperature applications such as underground engineering and resource development.

Applied Sciences doi: 10.3390/app16115371

Authors: Yao Li Faming Huang Yuqi Zheng Xiaomin Dai

Traffic flow prediction is important for route planning, signal control, and traffic guidance. However, traffic-state sequences usually exhibit non-stationarity, periodicity, and complex temporal dependencies, which makes it difficult for traditional statistical methods and single deep learning models to simultaneously capture short-term local fluctuations and long-term evolutionary trends. To address this issue, this paper proposes a dual-stream latefusion CNN-LSTM with an adaptive gated shortcut for traffic flow prediction, denoted as AGS-CNN-LSTM. The proposed method does not aim at explicit spatial-topology modeling; instead, it focuses on improving the fusion mechanism of CNN-LSTM-based models under settings without graph-structure constraints. Based on two public datasets, PeMS-BAY and PeMSD8, this study constructs multi-step prediction tasks with horizons of 15 min, 30 min, 60 min, 90 min, and 120 min and compares the proposed model with MLP, SimpleRNN, 1DCNN, LSTM, Serial CNN-LSTM, CNN-LSTM-Attention, BiLSTM-Attention, TCN-LSTM, Transformer Encoder, DLinear, and DS-CNN-LSTM (w/o Gate). The experimental results show that AGS-CNN-LSTM does not consistently achieve the best performance across all datasets, prediction horizons, and evaluation metrics. Nevertheless, it performs close to the best baseline models on the 30 min and 60 min tasks of PeMS-BAY and achieves competitive RMSE and R2 results on the 15 min, 30 min, and 60 min tasks of PeMSD8. Further ablation experiments indicate that the adaptive gated shortcut can enhance the predictive capability of the dual-stream late-fusion structure in some scenarios, although its benefits are dependent on the dataset and prediction horizon. Overall, the proposed model is more appropriately regarded as a lightweight fusion-mechanism improvement for CNN-LSTM-based models under settings without explicit graph-structure constraints, rather than a comprehensive replacement for complex graph neural networks, Transformerbased models, or models incorporating multiple external factors. Therefore, the findings should be interpreted as proof-of-concept evidence for a lightweight CNN-LSTM fusion enhancement under constrained non-graph-input settings, rather than as evidence of broad generalizability in complete road-network-level traffic forecasting.

Applied Sciences doi: 10.3390/app16115367

Authors: Sarangerel Jarantaibaatar Md. Shiful Islam Maximo Larry Lopez Caceres Yago Diez Myagmarjav Indra Tobias Leidemer Vladislav Bukin Shinsuke Konno Shinebayar Turbat Batbileg Bayaraa Jun Yokoyama Atsushi Nakamura Burmaa Chuluunbat Leonardo Huisacayna Silvestre Federico Giovanni Nicola

Weed infestation remains a major constraint in wheat production, highlighting the need for accurate and scalable monitoring approaches. Recent advances in unmanned aerial vehicle (UAV) and deep learning (DL) have created new opportunities for field-scale weed detection. This study evaluates the performance of three DL models (Mask R-CNN, YOLO26s-seg, and RT-DETR-L) for detecting common weed species in a 4-hectare rain-fed wheat field in Mongolia using UAV-acquired images. The following three weed species were identified as the most abundant under the dry environmental conditions of 2025: Linaria buriatica, Neneo pulla, and Artemisia scoparia. The models were evaluated using Precision, Recall, and F1-score. The detection F1-score for the L. buriatica, N. pulla and A. scoparia was 0.642, 0.568 and 0.163 for Mask R-CNN; 0.615, 0.655, 0.335 for YOLO26s-seg; and 0.693, 0.647, 0.275 for RT-DETR-L, respectively. Segmentation F1-score values were 0.659, 0.600, 0.179 for Mask R-CNN, whereas YOLO26s-seg achieved 0.613, 0.658, 0.326, respectively. RT-DETR-L achieved the best overall detection performance for the dominant species, while YOLO26s-seg showed the strongest detection of minority species, and Mask R-CNN produced the most accurate segmentation boundaries. Despite the imbalance in weed instances their morphological characteristics influenced the performance of the model. These findings demonstrate the feasibility of UAV imagery and DL models in precise weed management practices in Mongolia.

Applied Sciences doi: 10.3390/app16115369

Authors: Sławomir Dudziak Łukasz Czerwiński Jan Malanowski Mateusz Politański

This paper presents an extended combined experimental and numerical study on the vibration comfort assessment of a modern timber-framed public utility building. The research focuses on a lightweight skeleton floor system, representing a typical high-frequency floor. In situ vibration measurements were conducted under various walking excitations (single and multiple pedestrians) to determine key vibration parameters. Post-processing, which yielded root mean square accelerations and velocities, was performed using a custom-developed code in the Mathematica package. A finite element model was prepared in Dlubal RFEM 6 using shell and beam elements with offsets. The dynamic characteristics obtained from the FE modal analysis showed high consistency with the experimental data, with a relative error of approximately 5 % for the fundamental frequency. The vibration comfort was assessed using two distinct methodologies: the JRC report and the SCI P354 guide. Both approaches positively verified the floor’s vibration comfort, confirming its suitability for the intended use. The study demonstrates that the JRC methodology is more straightforward and unambiguous for engineering practice. Furthermore, the results indicate that simplified FE models provide a reliable basis for predicting vibration modes and calculating mode shape factors, which are essential for the correct interpretation of local measurements in existing buildings.

Applied Sciences doi: 10.3390/app16115368

Authors: Anusha Upadhyay Manob Jyoti Saikia

Comprehensive monitoring of cerebral hemodynamics has led to significant advances in Functional Near-Infrared Systems (fNIRS), particularly in terms of hardware design and development of wearable platforms. These advancements have established fNIRS devices as valuable tools in research and clinical practices; however, most existing literature focuses predominantly on clinical applications or high-level system performance. This review provides a rigorous, bottom-up analysis of bio-instrumentation architectures, evaluating the low-level trade-offs in component selection and circuit design that define modern wearable fNIRS performance. In this paper, we have identified and compared key hardware components of modern fNIRS technologies, including optical sensors, signal conditioning elements, control units, power systems, and communication modules. Significant progress has been made in terms of optical tomography, head coverage and conformity, multimodal integration, hyperscanning, motion tolerance, user comfort, and miniaturization. The paper underscores how systems may have unique architectures although they follow the same foundational principle. It also aims to identify the trade-offs existing in current fNIRS devices. Overall, this paper presents an overview of where we stand in terms of fNIRS development and attempts to trace an outline of the next generation of devices.

Applied Sciences doi: 10.3390/app16115366

Authors: Jurij Hladnik Franc Majdič Anže Čelik Boris Jerman

Preloaded threaded plugs that limit axial movement of spools in hydraulic valves are repeatedly loaded by spool impacts, which affects their fatigue strength. For proper dimensioning of these plugs, their loadings must be known. The aims of the study are to: (a) determine the spool impact force (Fimpact) on the plug head, (b) determine its influence on the force in the plug neck (FIPN) and (c) give design recommendations and practical considerations for plug–valve connections. Modified versions of the plugs were produced and equipped with strain gauges. The plugs were tightened into valves and loaded with spool impacts at three different pilot pressures. Meanwhile, Fimpact, FIPN, pilot pressure, and spool displacement were measured. As the FIPN was measured in the modified plug–valve connections, which had different stiffnesses from the original connections, the FIPN in the original plug–valve connections was calculated by considering the stiffness ratios of the original connections obtained from finite element analyses. The results show that the spool impact force and the resulting additional FIPN increase linearly with pilot pressure. To achieve smaller amplitudes of the FIPN and, consequently, higher fatigue strength, the tensioned parts of the plugs should have lower stiffness and the clamping parts higher stiffness.

Applied Sciences doi: 10.3390/app16115364

Authors: Artūras Serackis Mindaugas Jankauskas Anastasija Grubinskienė Vytautas Abromavičius

This study evaluates frame selection during inference as an independent factor in video deepfake detection while keeping the downstream detectors fixed. We compare twelve frame selection strategies, ranging from simple temporal and quality baselines to landmark aware policies, using four validated pretrained detectors: Self-Blended Images (SBIs), Frequency-Enhanced Self-Blended Images (FSBIs), Generative Convolutional Vision Transformer (GenConViT), and GenD. The primary experiment is a complete factorial benchmark with 300 videos and five frame budgets (2, 4, 8, 16, and 32 selected frames), which provides the reference results at 32 frames. To address sample size limitations, an additional validation experiment uses a deduplicated split of 1180 Celeb-DF++ and FaceForensics++ videos, with complete results for 2, 4, and 8 selected frames and a reported subset for 16 selected frames. In the complete 300-video benchmark, 32 frames achieved the strongest average AUC, while 8 and 16 frames recovered most of the attainable performance with lower runtime. The best single validated configuration was GenD with Shot-aware sampling at 32 frames, yielding an AUC of 0.9607 and a balanced accuracy of 0.9133. The study therefore does not claim that smaller budgets universally outperform 32 frames; instead, it quantifies the tradeoff between accuracy and runtime and shows that frame selection remains a meaningful design variable under constrained inference budgets.

Applied Sciences doi: 10.3390/app16115365

Authors: Izabela Rojek Jakub Kopowski Michał Rosiak Dariusz Mikołajewski

Composite materials, particularly polymers reinforced with carbon, glass, and aramid fibers, enable the development of lightweight yet mechanically robust structures that enhance user comfort and functional performance. Their high strength-to-weight ratio and fatigue resistance make them ideal for applications requiring repetitive movements in rehabilitation and assistive robotics. However, challenges remain related to cost-effective production, durability under complex loading conditions, and ergonomic fit to human anatomy. Recent advances in materials science and smart materials are expanding the possibilities of multifunctional composites with embedded sensors. Furthermore, machine learning methods are increasingly being used to optimize material selection and structural design. Future advances are expected to improve scalability, personalization, and system integration, positioning composite fibers as a key assistive technology in next-generation robotic systems.

Applied Sciences doi: 10.3390/app16115363

Authors: Guangmiao Huang Chengkang Lee Boyang Huang

This study evaluates a preliminary multi-objective optimization framework for a permanent magnet mechanical clutch designed for automated curtain actuators. To analyze the highly non-linear trade-off between disengagement capability (Y-direction magnetic resistance) and positional stability (Z-direction magnetic attraction force), a numerical approach combining three-dimensional (3D) magnetostatic finite element method (FEM) simulations, Taguchi L9 orthogonal arrays, and regression modeling was implemented. Three magnetic bead diameters were introduced as noise factors to investigate the sensitivity of the magnetic forces within a controlled simulation environment. A multiplicative composite objective function was employed to assess the competing performance criteria without masking single-factor failures. Statistical analysis indicates that within the investigated design space, the axial distance (Factor D) is the primary geometric parameter influencing the force distributions, followed by the outer diameter (Factor B) and inner diameter (Factor A). The identified parameter configuration (A = 8 mm, B = 10.5 mm, C = 0.5 mm, D = 1.5 mm) demonstrated an improved composite objective value and narrower standard deviations under the designated simulation boundaries compared to the initial discrete trials. These exploratory findings suggest that the proposed workflow was validated using a physical prototype based on one of the Taguchi configurations.

Applied Sciences doi: 10.3390/app16115362

Authors: Tzu-Yun Chang Zhen-Shu Liu Tzu-Heng Su Po-Wen Chen

Methodological variability remains a major source of uncertainty in environmental microplastics (MPs) analysis, particularly for micro-Fourier transform infrared (μ-FTIR) spectroscopy operated in reflection mode. This study quantitatively evaluates how key analytical procedures influence recovery efficiency and identification performance in μ-FTIR-based MPs analysis. Polystyrene (PS) standard particles (90 μm and 30 μm; density 1.05 g/cm3) were employed in direct titration and standard addition experiments. The effects of filter materials, rinsing suspensions, filtration approaches, and identification protocols were systematically assessed. Under the tested reflection-mode μ-FTIR conditions, silicon and stainless-steel filters provided sufficient spectral readability and microscopic particle visibility for PS particle recognition and were therefore selected for subsequent recovery evaluation. Rinsing with 50% ethanol reduced aggregation and enhanced recovery. Pump-assisted filtration (200 mmHg) achieved high PS recovery (88 ± 7.25%), and standard addition in river samples reached 91 ± 14.5%. Manual calibration using reference standards further improved classification consistency. These findings quantitatively link procedural choices to recovery and identification outcomes, providing practical guidance to improve reliability and inter-study comparability in μ-FTIR-based MPs analysis.

Applied Sciences doi: 10.3390/app16115361

Authors: Taejoon Kwon Si-Maek Lee Youngbin Kim Hyeongmin Yoo

Structural health monitoring of thick-walled industrial components remains challenging due to modal superposition and dispersion effects that limit conventional time-of-flight-based guided-wave analysis. This study proposes an active excitation-based monitoring framework using a self-developed flexible Macro Fiber Composite (MFC) transducer for defect characterization in 25 mm thick aluminum plates. Controlled three-cycle tone-burst excitation at 350 kHz was introduced, and the resulting elastic wave responses were analyzed using Short-Time Fourier Transform (STFT)-based time–frequency energy and spectral bandwidth metrics. Artificial V-shaped notches with depths of 10% and 30% of the plate thickness were introduced to evaluate defect severity. Compared to the intact specimen, the 10% notched plate exhibited a 22.5% reduction in STFT-based energy and a 3.37% decrease in spectral bandwidth, while the 30% notched specimen showed reductions of 37.5% and 7.78%, respectively. The results demonstrate that defect-induced structural discontinuities in thick plates not only attenuate overall guided-wave energy but also alter frequency distribution characteristics. The proposed approach enables quantitative defect evaluation without explicit modal separation and validates the dual actuation and sensing capability of the flexible MFC transducer, supporting the feasibility of transitioning from passive acoustic emission monitoring to an active, self-diagnostic structural health monitoring framework for thick industrial structures.

Applied Sciences doi: 10.3390/app16115360

Authors: Riku Yoshida Ryota Goka Keisuke Maeda Takahiro Ogawa Miki Haseyama

Composed Video Retrieval (CoVR) enables flexible video search by retrieving a target video that reflects a specified modification to a query video. The triplet datasets—consisting of query videos, query text, and target videos—required for model training have been collected manually. Recent studies have explored automatic construction of training triplets for CoVR; however, most existing approaches rely heavily on caption similarity. This limitation is particularly problematic in soccer videos, where identical or highly similar captions can correspond to visually distinct situations, making it difficult to construct triplets with appropriate relationships. To address this issue, this paper proposes a multimodal triplet construction framework specialized for soccer videos. The key idea is to explicitly incorporate visual similarity alongside textual similarity. Specifically, candidate target videos are selected by combining visual similarity with commentary caption filtering, enabling the identification of videos that are visually similar yet semantically different. The semantic difference between videos is then generated as query text using a large language model (LLM) without manual annotation. Furthermore, a multimodal large language model (MLLM) is introduced to estimate whether the generated modification is visually and semantically consistent with the video pair. Rather than replacing human verification, this step provides an automated screening signal to identify potentially unreliable triplets. The experiments show that the proposed framework automatically constructs triplets with reasonable validity under limited human validation. These results demonstrate the potential of scalable triplet construction for CoVR in soccer videos.

Applied Sciences doi: 10.3390/app16115359

Authors: Zhadyra Alimbayeva Chingiz Alimbayev Kassymbek Ozhikenov Aiman Ozhikenova Ussen Shylmyrza Kymbat Khaidarova

This structured critical review provides a comparative and analytically grounded overview of machine learning (ML) approaches for electrocardiography (ECG)-based detection of dysglycemia, with a specific focus on translational readiness for clinical screening. A structured literature search across PubMed, Scopus, Web of Science, and IEEE Xplore identified 183 records, of which 17 studies were included following predefined screening criteria and PRISMA-guided selection principles. The included studies demonstrate substantial heterogeneity in dataset size (ranging from <50 to >25,000 subjects), ECG acquisition modalities (single-lead, 12-lead, wearable), feature representations (raw signals, heart rate variability, engineered features), and ML strategies (classical algorithms, deep learning, and multimodal models). Reported model performance is generally high, with accuracy values frequently exceeding 0.85 and area under the curve (AUC) ranging from 0.78 to 0.99. Smaller experimental studies often report inflated performance (up to 96–99% accuracy), whereas large-scale population-based investigations demonstrate more moderate but clinically plausible results (AUC ≈ 0.80–0.85). External validation, a key requirement for clinical applicability, was performed in only a limited subset of studies (approximately 12%). From a physiological perspective, ML models exploit ECG alterations associated with dysglycemia, including reduced heart rate variability, QT interval prolongation, and changes in ventricular depolarization and repolarization dynamics. However, the relationship between metabolic dysfunction and ECG signals remains indirect. A key finding of this review is the mismatch between reported predictive performance and translational readiness. The majority of studies (≈65–70%) are classified as early-stage (Level 1–2 or 2–3), relying on small, single-center datasets and internal validation. Only a minority of studies achieve near-translational maturity (Level 4), characterized by large-scale datasets and external validation. ECG-based dysglycemia detection represents a promising non-invasive and scalable screening paradigm. However, its clinical translation is constrained by the lack of standardized ECG acquisition protocols, limited dataset diversity, insufficient external validation, and fragmented methodological approaches. Future research should prioritize large multi-center datasets, standardized feature extraction pipelines, hybrid interpretable models, and prospective validation to enable robust, generalizable, and clinically deployable screening systems.

Applied Sciences doi: 10.3390/app16115358

Authors: Abdullah Faiz Kun Li Ping Chen

Underwater image analysis is affected by light scattering, wavelength-dependent attenuation, low contrast, and suspended particles, which reduce the discriminative visual features. Current multi-scale Vision Transformers are not well-suited to these degradations because they cannot effectively fuse features across scales to achieve accurate classification. Although Vision Transformers (ViTs) can model long-range interactions, single-scale patch tokenization remains suboptimal for underwater images, where both fine-grained textures and global structures are important. This study proposes a Multi-Scale Vision Transformer (MS-ViT) with Cross-Scale Biased Attention Fusion (CSBAF) for underwater image classification. Before transformer encoding, the CSBAF introduces a learnable source–target scale-pair bias and an input-dependent scale-reliability gate. This differs from standard multi-scale fusion and cross-attention methods, which mainly concatenate features or exchange information between scale branches. The proposed design enables the model to emphasize reliable scales while suppressing degraded-scale responses. A hybrid dataset containing 14,000 images from the Roboflow Aquarium and RUIE datasets across five classes was used for evaluation. MS-ViT with CSBAF achieved 88.9% accuracy and an 88.8% F1-score, outperforming the CNN baseline by 7.6% and state-of-the-art transformer models, including UWFormer, DP-ViT, and CvT, by 2.3–4.2%. Ablation studies showed a 1.7% accuracy improvement over simple multi-scale concatenation, whereas cross-dataset testing achieved 84.4% accuracy, indicating reasonable cross-dataset robustness. These results demonstrate that explicit scale–aware fusion can improve transformer-based underwater visual understanding.

Applied Sciences doi: 10.3390/app16115357

Authors: Cecília Bérgamo Biancardi André Silva de Carvalho

Tile adhesive mortars are industrialized products used for installing ceramic coverings and are classified according to the Brazilian standard ABNT NBR 14081/2012 on the basis of tensile adhesion performance under different curing conditions. Their formulation directly affects both technical performance and manufacturing competitiveness, while conventional product development remains slow, costly and strongly dependent on trial-and-error laboratory testing. This study evaluates whether historical industrial formulation data can support the retrospective prediction of approval or failure of tile adhesive mortars under ambient, oven, immersed and open-time curing conditions. A dataset comprising 6031 individual pull-off observations collected between 2021 and 2023 by a European multinational company in the construction materials sector was used to train and compare Logistic Regression, Random Forest, Boosted Decision Tree and Support Vector Machine models in R and Azure. The study was designed as an industrial-data modelling investigation rather than as a prospective optimization experiment. The results show that ensemble tree-based models, particularly Boosted Decision Tree and Random Forest, achieved the strongest predictive performance, whereas Logistic Regression remained more suitable for inferential interpretation of formulation variables. Model performance was uneven across curing conditions: prediction was more reliable for oven and immersed curing, whereas ambient curing and open time were affected by strong class imbalance and low failure prevalence. The findings indicate that Machine Learning can support formulation screening and quality-oriented decision-making for tile adhesive mortars, provided that its use remains restricted to the formulation ranges represented in the historical dataset and is complemented by prospective experimental validation before deployment in new product development.

Applied Sciences doi: 10.3390/app16115356

Authors: Marco F. L. Lemos Susana F. J. Silva Ana Augusto

Marine-derived compounds are increasingly being reported as promising candidates for use in food products due to their antioxidant, antimicrobial, and other bioactive properties. However, the successful translation of these compounds into effective food ingredients lags far behind the growing body of bioactivity data. This discrepancy reflects the tendency to equate activity measured under simplified laboratory conditions with functionality in real food systems. This article argues that such an assumption is often misleading. The performance of marine bioactives in food matrices is affected by factors such as instability, interactions with surrounding components, processing conditions, and loss of efficacy over time. Consequently, conventional in vitro screening often overestimates application potential and has limited predictive value for practical use. To advance the field, we propose a functionality-driven translation framework that shifts the evaluation focus away from bioactivity-centred screening towards assessing stability, matrix compatibility, feasible dosages and performance under conditions mimicking food matrix complexity. Better alignment between discovery and application is essential if the diversity of marine chemicals is to generate robust and effective solutions for food systems.

Applied Sciences doi: 10.3390/app16115355

Authors: Cornel Ilinca

The sustainable management of water resources has become a fundamental necessity for modern society, driven by the intensifying pressures of industrialization, rapid population growth, and the overarching impacts of climate change [...]

Applied Sciences doi: 10.3390/app16115354

Authors: Yanming Chen Qiming Sun Ying Zhang Chengxuan Li

Multi-agent systems are widely used in modern power systems, but they face challenges such as low data utilization, stringent triggering conditions, and poor environmental adaptability. This study proposes a multi-agent event-triggered control method based on the Proximal Policy Optimization (PPO) policy gradient algorithm. By maximizing the cumulative reward, the agents are driven to learn adaptive triggering strategies, which reduces communication frequency while ensuring system stability. A multi-agent reinforcement learning model is constructed, and the training results show that both the single-episode reward and the average reward significantly increase with the number of training episodes, thus verifying the effectiveness of the algorithm. Based on Lyapunov stability and LaSalle’s invariance principle, an event-triggering threshold is designed using an exponential decay function. Moreover, the sequential decision-making process under uncertain environments is described using the Markov decision process. In the case study with six agents, the triggering conditions effectively constrain the error growth and ensure system stability. The method is further extended to a 33-node power system, where each node is regarded as an agent to simulate voltage fluctuations under load variations. Compared with periodic sampling control, the event-triggered control exhibits faster convergence speed, higher steady-state accuracy, and stronger anti-interference capability, thus confirming its superiority in complex power systems.

Applied Sciences doi: 10.3390/app16115353

Authors: Haoxuan Xiao Jianfeng Zheng

To address the challenges of limited samples, class imbalance, and real-time requirements in surface defect detection for solar vacuum glass collector tubes, this study proposes an improved lightweight RT-DETR-based method. Specifically, DICM is introduced into the backbone to improve multi-directional and multi-scale feature extraction, HAFB is embedded in the neck to enhance the fusion of local details and global semantics, and transfer learning is adopted to alleviate data scarcity under small-sample conditions. Experiments on a self-built defect dataset of solar vacuum glass collector tubes show that the proposed method outperforms the original RT-DETR and several mainstream detectors in terms of Precision, Recall, mAP@0.5, and F1-score while maintaining favorable inference speed and model compactness. Under the same hardware conditions, the proposed model achieves an mAP@0.5 of 0.95, an inference speed of 83.21 FPS, and a model size of 82.36 MB. These results demonstrate the feasibility of the proposed method for real-time online defect detection in industrial scenarios.

Applied Sciences doi: 10.3390/app16115352

Authors: Ahmed M. Ahmed Jeffrey Shragge Ilya Tsvankin

Real-time seismic reservoir monitoring of geologic reservoirs requires both large-scale data management and efficient computational workflows. Addressing these challenges is facilitated by developing techniques capable of selectively capturing critical geologic features, thereby increasing computational efficiency and reducing data storage requirements. This paper proposes a double-compression framework integrating Haar wavelet transforms with machine learning (ML) for efficient multiparameter seismic inversion. First, Haar wavelet compression significantly reduces the dimensionality of the input elastic models, preserving essential geologic structures while limiting data volumes. Next, a convolutional neural network with the long short-term memory (CNN-LSTM) architecture, including dual encoders and multi-decoders, compresses seismic data into a latent space to generate a multi-scale P-wave velocity estimate. By leveraging transfer learning to speed up convergence and enhance prediction accuracy, we fine-tune the latent representation to estimate the P-to-S-wave velocity ratio and acoustic impedance at multiple resolution scales. Tests on the synthetic CO2-injection Kimberlina model show that wavelet-based compression—including detuning large-scale trends—minimizes artifacts in simulated wavefields and accelerates neural-network training. The results demonstrate that combining wavelet-based pre-compression for reservoir models with data-driven latent encodings for seismic data achieves high compression ratios, reduces computational costs, and maintains the fidelity of subsurface imaging. Compared with a redundant-decimation baseline, the proposed framework reduces network training time by approximately 70% and GPU memory usage by 33–73%, achieves a wavefield energy loss below 0.1% at a 16:1 model-dimension reduction, and produces multi-resolution predictions of VP, VP/VS, and acoustic impedance with normalized errors below 0.04 across all six wavelet decomposition levels. Thus, the double-compression framework enables robust and scalable seismic monitoring of elastic reservoir parameters.

Applied Sciences doi: 10.3390/app16115351

Authors: Marui Han Zhiping Zeng Zijie Li Peicheng Li Guangzhao Peng Weidong Wang Abdulmumin Ahmed Shuaibu

The wheel–rail impact effect is prominent in the low vibration track (LVT) in the curved sections of heavy-haul railways, where fastener failure and the support block hanging void are prone to occurring. To investigate the impact of these issues on the dynamic characteristics of the vehicle–track coupled system, this study establishes a coupled dynamics model of a heavy-haul train and LVT, taking into account the topological relationships of vehicle components, multipoint wheel–rail contact, and track irregularities. Comparative analyses are conducted to evaluate the effects of the location, quantity, and failure degree of fastener failure and support block hanging voids on running safety and stability. The results show that (1) compared to the normal condition, fastener failure and support block hanging voids lead to varying degrees of increases in response indicators, thereby intensifying the wheel–rail impact; (2) bilateral failure exhibits more pronounced dynamic responses than unilateral failure, and when the number of failed fasteners or hanging voids exceeds one, the maximum wheel load reduction rate increases significantly; (3) as the gap of the hanging void increases, the dynamic response also increases, and when the gap reaches approximately 3 mm, the support block can be considered fully suspended; and (4) comprehensive analysis indicates that fastener failure poses a greater threat to running safety than support block hanging voids and thus warrants greater attention in practical engineering applications. This study provides theoretical support for the maintenance and repair of heavy-haul railways.

Applied Sciences doi: 10.3390/app16115350

Authors: Gaspar Cano José Manuel Ortega-Candel Francisco José Mora-Gimeno Lucía Arnau-Muñoz Higinio Mora

In the era of cloud computing and serverless architectures, the security of applications and services has become a critical challenge. Serverless computing, often referred to as function as a service (FaaS), is a cloud computing model that allows developers to build and run applications without the need to manage traditional server infrastructure. Serverless architectures have gained popularity in cloud computing due to their flexibility and ability to scale automatically based on demand. These architectures are based on executing functions without the need to manage the underlying infrastructure. Denial of wallet (DoW) attacks refer to a type of cyberattack that aims to exploit and exhaust the financial resources of an organization by triggering excessive costs or charges within their cloud or serverless computing environment, exploiting characteristics such as the pay-as-you-go model, auto-scaling, limited control, and cost amplification. This research aims to assess existing methods for detecting distributed denial of service (DDoS) attacks and extend their application to detect denial of wallet (DoW) threats, leveraging a dataset tailored to serverless architectures. We investigate various strategies and techniques that employ entropy, machine learning and deep learning algorithms to enable early detection of DDoS and DoW attacks in serverless environments. This research provides insights into the options that are available for detecting DoW attacks in serverless environments, allowing security professionals and developers to make decisions on the most appropriate solutions to protect their applications and cloud services.

Applied Sciences doi: 10.3390/app16115349

Authors: Rendra Hakim Hafyan Vinod Kumar Sunil K. Maity Jhuma Sadhukhan Siddharth Gadkari

Food waste (FW) is a major global challenge with significant economic and environmental costs, yet its nutrient-rich composition also offers an opportunity for valorization into high-value biochemicals and biofuels within a circular bioeconomy. Effective FW management requires systematic frameworks that balance environmental performance, economic returns, and social acceptance, a challenge that is particularly difficult in developing countries where technical, financial, and participation barriers persist. This review proposes a strategic, step-by-step approach to enhance current FW management through the objective integration of biorefinery pathways producing biochemicals and biofuels products. Both biochemical and thermochemical conversion routes are evaluated against industrial yield benchmarks, market value, and end-use specifications to identify the products and processes most capable of enhancing sustainability. The review further presents a framework for multi-objective optimization (MOO) that simultaneously addresses economic, environmental, and social objectives, and for incorporating decision-maker preferences into the selection of optimum solutions. By coupling sustainability assessment with structured decision support, this review provides practical guidance for selecting FW management strategies that are economically viable, environmentally sound, and socially acceptable.

Applied Sciences doi: 10.3390/app16115345

Authors: Ziemowit Bańkosz Pengfei Jin Anna Węgrzyn Katarzyna Węgrzyn Sławomir Winiarski

(1) Background: The aim of this study was to determine the biomechanical role of plantar pressure distribution in generating racket velocity during the topspin forehand in table tennis players, with particular emphasis on its relationship with stroke kinematics and performance level. (2) Methods: The study involved 14 male table tennis players divided into Elite and Sub-Elite athletes. Each participant performed a topspin forehand stroke. The study employed a biomechanical analysis combining inertial motion capture and plantar pressure measurement to assess the relationship between lower limb loading and racket velocity during the topspin forehand. (3) Results: The statistical evidence supports the subsequent phase-by-phase comparisons, indicating that the Elite (EL) and Sub-Elite players (SE) differ in execution of the topspin forehand, and the Elite group achieved significantly higher racket speed values in all phases (e.g., in hitting phase: SE-13.8 m/s, EL-15.6 m/s, p ≤ 0.001, d = 1.0; in post-impact follow-through phase: SE-13.8 m/s, El-16.1 m/s, ≤0.001, d = 1.3) and exhibited also a different pattern of foot loading. An analysis of the correlation between the plantar pressure and velocity of the racket in individual events revealed numerous significant correlations. (4) Conclusions: The study identified numerous correlations between the maximum plantar pressure and the maximum racket speed in the individual phases of the stroke. This demonstrates the active involvement of the feet throughout the entire kinematic chain of the topspin forehand stroke and highlights the importance of foot coordination for the outcome of this stroke, namely the speed of the racket-wielding arm.

Applied Sciences doi: 10.3390/app16115348

Authors: Zhimeng Chen Kejia Hu Junqiang Liu Yinkai Ji Yi Zhu Hualun Chen Chao Yuan Zhiyu Chen

Underground duct cables are widely used in urban power distribution systems, but their enclosed installation environment makes defect inspection difficult, labor-intensive, and potentially hazardous. Infrared thermography can capture abnormal temperature distributions caused by insulation degradation, conductor damage, sheath failure, or severe structural defects, while robot-based inspection provides a promising solution for confined duct environments. However, thermographic fault detection for underground small-diameter duct cables remains insufficiently studied, and practical deployment requires lightweight models suitable for embedded edge devices. In this study, an improved YOLO26-based thermographic fault detection framework is proposed for underground duct cable inspection. A Cable-Thermo dataset is constructed using an ANSYS 2025 R2-based thermoelectric coupling simulation, covering four defect categories: hollow-type damage, conductor burnout, sheath damage, and severe damage. To balance detection accuracy and deployment efficiency, two model variants are developed. YOLO26-Thermo-E retains the original detection scales and integrates CDA and SimSPPF modules for accuracy-prioritized diagnosis. YOLO26-Thermo-H further removes the small-scale detection branch as a deployment-oriented design choice, based on the scale distribution observed in the simulation dataset, where most fault-induced thermal anomalies appear as spatially continuous medium- or large-scale regions. This design assumption still requires further validation using real duct thermographic data. Experiments show that YOLO26-Thermo-E achieves the highest mAP50 of 99.20%. YOLO26-Thermo-H maintains a mAP50 of 99.00% while reducing GFLOPs by 34.3% and parameters by 16.2% compared with YOLO26. On an NVIDIA Jetson Orin NX, YOLO26-Thermo-H reaches 34 FPS under FP16 inference and 45 FPS under INT8 inference. These results demonstrate the feasibility of the proposed framework under controlled simulation conditions and its potential for edge deployment. The limitations of the simulation-based dataset are also discussed, and future work will focus on real-scene data collection and simulation-to-real generalization.

Applied Sciences doi: 10.3390/app16115347

Authors: Mieczysław Cichoń Kinga Cichoń

This paper presents a generalized epidemic modeling framework based on g-tempered Caputo fractional derivatives with discrete time delays. The proposed approach incorporates nonlocal memory effects, nonlinear temporal scaling, and delayed epidemiological responses within a unified mathematical structure. The introduction of the nonlinear time transformation g(t) and the tempering parameter λ eliminates the unrealistic infinite-memory behavior associated with classical power-law kernels while simultaneously introducing new challenges related to parameter identifiability and inverse problems. We investigate the structural properties of the resulting dynamical systems and show that the associated inverse problem is inherently ill-posed. To illustrate the practical implications of these results, the framework is applied to a delayed SIQR epidemiological model. Numerical simulations are performed using a generalized L1-type scheme adapted to delayed fractional histories, and a multi-phase parameter estimation procedure is proposed to address the ill-posedness of the reconstruction problem. The results demonstrate the ability of the model to capture both short- and long-term memory effects in epidemic evolution while highlighting the challenges of statistical identifiability in generalized fractional systems.

Applied Sciences doi: 10.3390/app16115346

Authors: Fran Oficial-Casado Alexis Soriano-Gandia Jose Ignacio Priego-Quesada

Background: Cycling performance depends on both aerobic capacity and neuromuscular function, with recent training approaches emphasizing the role of strength training. However, the extent to which neuromuscular characteristics assessed in conventional strength exercises transfer to cycling performance remains unclear. Therefore, the aim of this study was to analyze the relationship between the Load–Velocity (L-V) profile obtained from a multi-joint strength exercise (half squat) and the cycling Power Profile (PP) in Masters-level cyclists. Methods: Twelve masters-level cyclists were evaluated by the L-V and the PP test. The cycling PP was determined through maximal efforts of 1, 5, and 20 min, expressed relative to body mass (W·kg−1). The L-V profile was assessed during the half squat using a progressive loading protocol with load–velocity monitoring. Pearson’s correlation analyses were performed between the slope and intercept of the L-V profile relationship and PP variables, as well as mean ascent velocity (VAM). Results: No significant relationships were observed between L-V profile variables and cycling performance (r = −0.21 to 0.09, p > 0.05). In contrast, VAM showed very large associations with P1 (r = 0.81, p = 0.001) and P5 (r = 0.86, p < 0.001). The regression model explained a large proportion of the variance in VAM (R2 = 0.75, p = 0.01). Conclusions: Strength performance assessed through a conventional exercise such as the half squat is not directly related to cycling PP in masters-level cyclists. The observed relationships between VAM and cycling PP reinforce the importance of task specificity.

Applied Sciences doi: 10.3390/app16115344

Authors: Ilare Bordeasu Dorin Bordeasu Brandusa Ghiban Daniel-Catalin Stroita Liviu-Daniel Pirvulescu Imre Kiss

The degradation of material surface structures caused by cavitation is of a mechanical nature due to cyclic local fatigue loading. Its intensity depends on both the material’s mechanical properties and its surface microstructure. Evaluating the surface structure resistance to cavitation loading can be performed based on experimentally determined mechanical properties and/or through macro- or microscopic analysis of the eroded structure. Manufacturers, designers, and users of hydromechanical equipment operating under cavitation conditions are interested in materials whose properties and structures can withstand cavitation loading. For this reason, the current experimental research in the field focuses on establishing relationships that express the influence of either mechanical properties or surface microstructure on the resistance of material structures to cavitation erosion. The current paper aligns with this research direction and aims to determine statistical relationships between mechanical properties and the cavitation erosion resistance of aluminum-based alloys. The mechanical properties considered include ultimate tensile strength (Rm), yield strength (Rp0.2), surface hardness (HB), resilience (KCU), and elongation at fracture (A5). Cavitation resistance is evaluated using the parameter Rcav, defined according to the ASTM G32-2016 standard. The experimental results were obtained from cavitation tests conducted using a standard vibratory device that complies with ASTM G32 requirements.

Applied Sciences doi: 10.3390/app16115343

Authors: Zhiqi Yang Hao Zhang Zhigang Xu Shihong Ge

This paper addresses an integrated production–transportation scheduling problem in assembly workshops, encompassing the processes of part machining, material handling via handling resources, and final synchronous assembly. The finite buffer capacities of production resources can cause blocking, thereby reducing efficiency. Material handling resources are subject to different service area restrictions, and some share safety zones with production resources, preventing simultaneous processing. To address this, a mixed-integer programming model is formulated with makespan and total empty travel time as bi-objective optimization targets. Since the mixed-integer linear programming (MILP) model faces difficulties in solving medium- and large-scale instances, an improved memetic NSGA-II algorithm (IMNSGA-II) is proposed. The algorithm adopts a three-segment chromosome encoding and incorporates a VNS-SA local search mechanism within the global evolutionary framework of NSGA-II. Small-scale computational experiments using Gurobi are first used to verify the correctness of the model. Decoupling experiments further demonstrate the necessity of integrated optimization: compared with phased baseline methods, IMNSGA-II reduces makespan and empty travel time by approximately 10.16% and 12.33%, respectively. In ablation and comparative experiments, results based on hypervolume (HV) and inverted generational distance (IGD) show that the proposed method achieves better convergence, diversity, and overall Pareto front quality than multiple baseline algorithms. These experiments confirm the effectiveness of the proposed model and algorithm.

Applied Sciences doi: 10.3390/app16115342

Authors: Alberto Borrega-Solano Pablo Lopez-Sierra Amalia Campos-Redondo Javier Garcia-Rubio

The present study examined whether offensive play-type indicators in professional basketball reflect broader latent playing-style dimensions and whether play-type usage or efficiency is more strongly associated with competitive success. Data were obtained from the official NBA statistics website and included 6400 games across five seasons (2019–2020 to 2023–2024), comprising 5979 regular-season games and 421 playoff games. For each offensive play type, two indicators were analysed separately: usage percentage and efficiency, operationalised as points per possession (PPP). Principal component analyses were conducted independently for regular-season and playoff data, and for usage and efficiency variables. In addition, linear mixed-effects models were used to examine the relationship between play-type indicators and competitive success while accounting for games nested within teams. Only regular-season efficiency variables showed adequate sampling adequacy for factorial analysis (KMO = 0.774), yielding a four-component solution that explained 58.85% of the total variance. In the mixed-effects models, usage variables were not significantly associated with success, whereas efficiency indicators showed greater explanatory value. Specifically, pick-and-roll ball handler PPP and spot-up PPP emerged as the strongest positive predictors of success, with smaller effects observed for roll-man PPP and cut PPP. The efficiency-only model improved model fit relative to the frequency-only model (marginal R2 = 0.799 vs. 0.755), whereas adding usage variables to efficiency provided only a negligible additional contribution (marginal R2 = 0.803). These findings suggest that, in the NBA, competitive success is more closely related to the effectiveness with which offensive actions are executed than to the relative frequency with which they are used. From an applied perspective, play-type efficiency appears to provide more actionable information than usage-based summaries for performance analysis and tactical decision-making.