As Ho Chi Minh City, Vietnam continues to develop its trade, manufacturing, port, and transportation sectors, the demand for effective and sustainable logistics has increased greatly. In this study, an integrated decision support framework with Covariance-Based Structural Equation Modeling (CB-SEM), Decision-Making Trial and Evaluation Laboratory (DEMATEL), Stepwise Weight Assessment Ratio Analysis (SWARA), and VlseKriterijumska Optimizacija I Kompromisno Resenje (VIKOR) was proposed to identify, prioritize, and evaluate the critical determinants for logistics service development in Ho Chi Minh City. Primary data were collected from 475 logistics-related enterprises across 9 latent constructs and 38 evaluation sub-criteria. The SEM results show that Technology (β = 0.234), Logistics Infrastructure (β = 0.196), Firm Capabilities (β = 0.189), and Capital and Investment (β = 0.186) have the strongest positive impact on logistics development (p < 0.001). Based on the DEMATEL analysis results, Technology (D+R = 5.5179) is the most prominent, and Regional Linkages (D−R = 0.6976), Market Demand (0.5953), and Green Logistics (0.3961) are the main causal factors. Enterprise software deployment (0.039), supply chain data integration (0.038), cloud/IoT adoption (0.037), and warehouse automation (0.037) are the most influential sub-criteria in the integrated SEM-SWARA framework. Long-term forecasting also shows logistics demand indices with low, base, and high growth of 330, 626, and 966 by 2045, respectively, indicating a need for strategic expansion of logistics infrastructure, digital transformation, and a sustainable logistics plan.
Maternal–fetal attachment (MFA) is the emotional bond between a pregnant woman and her fetus and is an important determinant of pregnancy outcomes. This study aimed to explore the experiences of pregnant women in the third trimester regarding MFA activities. This qualitative study used a phenomenological approach and was conducted at a public health service in Yogyakarta, Indonesia, which does not yet have a specific MFA program reflecting unique regional cultural characteristics. We purposively recruited seven pregnant women with normal pregnancies at 35–40 weeks of gestation. We collected data through semi-structured interviews and analyzed them using thematic analysis. The MFA activities carried out by the participants were referred to as the Indonesian Prenatal Attachment Inventory instrument. This study identified five themes: understanding of MFA, MFA activities, factors that support and hinder MFA, and the role of midwives in MFA. A distinctive finding was the strong influence of cultural and spiritual values on MFA. Participants viewed their unborn children as divine trust and perceived MFA activities not only as instinctive but also as a moral and spiritual responsibility. Factors supporting MFA included previous pregnancy experience, gestational age, social support, and cultural and spiritual beliefs. Barriers to MFA included unplanned pregnancy, maternal health, multiparity, and miscarriage. These findings highlight the need for midwives to incorporate biological, psychological, social, spiritual, and cultural dimensions into antenatal care to strengthen MFA and maternal, neonatal, and child health outcomes.
Hospital logistic systems manage the flows of materials and information that support patient care, with pharmacy management playing a central role in both patient safety and operational efficiency. High rates of medication errors and excessive staff workload are persistent challenges in hospital pharmacies, which automation technologies may help address. This study evaluates the operational impact of integrating automatic unit-dose dispensers, autonomous mobile robots (AMRs), and personal medication dispensers (PMDs) in medium-sized hospital settings. Using discrete-event simulation (DES), we compared two technology-driven scenarios with a conventional baseline. In Scenario 1, Centralized Automation, an automated dispenser in the central pharmacy prepares medications as unit doses. In Scenario 2, Ward-level Automation, drugs are manually prepared in daily doses at the pharmacy and subsequently dispensed at the bedside using PMDs. The simulation results indicate that both approaches reduce the workload of transcription staff, dispensing staff, and pharmacists by 10–17%. In particular, Scenario 1 achieves the largest reduction, reducing dispensing staff workload by 17%. Furthermore, automation improves the consistency and safety of medication delivery. These findings provide empirical support for adopting integrated automation systems as a strategic approach to enhance workflow efficiency, reduce human error, and improve patient safety in hospital pharmacy operations. This study demonstrates that carefully planned technology integration can deliver measurable operational benefits while maintaining high standards of care.
Predictive maintenance in smart factories requires not only high prediction accuracy but also low-latency processing, adaptability to data drift, and understandable explanations for field operators. However, many existing approaches remain cloud-centered, label-dependent, and weak in practical explainability. This study proposes an explainable predictive maintenance framework based on FPGA-GPU Edge–Cloud hybrid computing for large-scale multivariate time-series environments. In the proposed system, FPGA modules perform streaming-oriented signal preprocessing and low-latency feature extraction, while GPU modules execute deep learning-based anomaly detection and fault prediction. To reduce dependence on labeled fault data, the framework incorporates masked autoencoder-based self-supervised representation learning. To improve long-term robustness in changing manufacturing environments, the framework also considers continual learning based on Elastic Weight Consolidation. In addition, a lightweight large language model with parameter-efficient fine-tuning and retrieval-augmented generation generates root-cause-oriented explanations and maintenance guidance. The method is organized as an integrated pipeline that combines data acquisition, edge preprocessing, temporal inference, explanation generation, and cloud-assisted model adaptation. The evaluation framework includes certification-oriented testing, comparative analysis, latency and throughput analysis, drift response analysis, and explainability assessment. According to a third-party test report, the proposed system achieved an event recall of 0.9822, event precision of 0.9529, event F1-score of 0.9674, and a false alarm rate of 0.000205. These results indicate that the proposed framework is practically feasible for real-time, explainable, and deployable predictive maintenance in smart factory environments.
Inventory management in multi-echelon supply chains plays a central role in improving efficiency and profitability. However, it remains complex because demand variability, lead time fluctuations, and supply disruptions affect multiple stages. Although the field is supported by quantitative models, stochastic approaches, and digital technologies, current studies still show limited integration of sustainability, emerging digital tools, and comprehensive risk assessments in dynamic multi-level systems. This paper evaluates research performance on inventory uncertainty in multi-echelon supply chains and identifies major trends, research gaps, and future directions. The study uses a bibliometric approach and analyzes 100 Scopus-indexed publications collected on November 13, 2024, including forthcoming 2025 articles. The review follows the PRISMA framework for eligibility, screening, and inclusion, with keywords related to inventory uncertainty and multi-echelon systems. Data were processed through Biblioshiny in R for Scientometric analysis, including visualizations of source relevance, author productivity, institutional output, country contributions, citation patterns, keyword co-occurrences, and thematic evolution. The results show strong growth in publications after 2014, with significant contributions from Iranian universities and applications in the food industry. Leading journals include Computers and Industrial Engineering, which shows high relevance and impact. Dominant themes focus on supply chains, stochastic systems, inventory control, and optimization. Emerging interests include uncertainty analysis and reinforcement learning. Research gaps remain in integrating risk management with blockchain, predictive analytics, and sustainability across supply chain tiers. This study highlights the need for adaptive inventory strategies that help managers strengthen resilience and support new research on innovative models for complex uncertainty.
The adoption of big data analytics (BDA) and artificial intelligence (AI) in hospitals is widely assumed to follow different logics across ownership types and service tiers, yet this heterogeneity is seldom tested formally. This study examines whether an integrated Resource-Based View and Technology Acceptance Model (RBV-TAM) of organizational readiness operates uniformly across Indonesian hospitals, and identifies which capabilities warrant managerial priority. Data from 330 hospital key informants spanning 128 public and private institutions across 24 provinces were modeled using partial least squares structural equation modeling. Measurement equivalence was assessed using the measurement invariance of composite models (MICOM) procedure; structural equivalence using permutation-based and Henseler multi-group analysis (MGA) across two managerially salient partitions, namely public versus private ownership and advanced versus basic service complexity; and managerial priorities using importance-performance map analysis (IPMA). Partial measurement invariance held for both partitions, and no structural path differed significantly across groups (ownership p = 0.230 to 0.237; complexity p = 0.597 to 0.626), indicating a homogeneous and generalizable adoption mechanism. Advanced hospitals scored higher on every latent construct, revealing a same-mechanism, different-endowments pattern rather than divergent behavior. IPMA identified perceived ease of use as the principal leverage point, combining above-average importance with the lowest performance, whereas behavioral intention was already strong. The findings support a universal readiness model and prioritize usability interventions and resource equalization for lower-tier hospitals. Further research should test the model longitudinally and in other health systems to confirm its cross-context stability.
Electrical equipment maintenance at Jakarta LRT stations currently relies heavily on manual inspections, which are time-consuming, labor-intensive, and may hinder early fault detection. This study research designs and implements an IoT-based monitoring system for continuous, real-time observation of critical electrical parameters to enhance maintenance efficiency and reliability. The system integrates a Digital Power Meter to monitor voltage, current, power, and energy consumption, alongside a Digital Temperature Controller with multiple sensors for thermal monitoring. Data is communicated via Modbus RS-485 and transmitted to a web-based Node-RED dashboard using MQTT. This centralized platform allows remote, user-friendly monitoring by operators. Experimental results demonstrate that the developed system achieves a measurement deviation of less than 2% relative to standard reference instruments, indicating reliable accuracy and stability. Furthermore, the inspection duration was significantly reduced from an average of 30 minutes to approximately 5 minutes per inspection cycle, while the required number of personnel decreased from 2 to 1. These improvements highlight the system’s contribution to operational efficiency, reduced resource utilization, enhanced situational awareness, and faster response to potential equipment failures. Beyond its direct application at Jakarta LRT stations, this research provides a practical and scalable reference for the broader adoption of IoT-based monitoring systems in transportation and energy infrastructure. Future developments may integrate predictive analytics and machine learning techniques to support failure prediction and data-driven preventive maintenance strategies.
Vietnam's rapid industrialization and urbanization have driven unprecedented energy demand growth, posing challenges to energy security, sustainability, and SDG 7 compliance amid rising carbon dioxide (CO₂) emissions. Vietnam has experienced rapid growth in energy consumption over the past few decades, driven by industrialization, urbanization, and rising living standards. Vietnam relies heavily on fossil fuels such as coal and oil, which have contributed to increasing CO₂ emissions. As a result, Vietnam is now one of the fastest-growing sources of greenhouse gas emissions in Southeast Asia. Therefore, this study addresses critical gaps in long-term energy forecasting for developing economies by applying Seasonal Autoregressive Integrated Moving Average (SARIMA) models to annual energy consumption data (1995-2023) sourced from the World Bank, International Energy Agency, and Vietnam's General Statistics Office. High correlations (0.94-0.99) among population, GDP, energy use, and emissions underscore their interdependence, validating time-series approaches for univariate prediction. Optimal SARIMA models project a steady escalation in energy demand from 206.76 TWh (2024) to 253.70 TWh (2030), with narrow confidence intervals indicating predictive stability. CO2 forecasts reveal volatility, peaking at 503.59 Mt (2029) before declining, highlighting model sensitivity to historical shocks. Comprehensive diagnostics, including residuals, ACF plots, and rolling means, confirm a robust fit for energy trends but residual autocorrelation in emissions, suggesting the need for hybrid enhancements. These projections equip policymakers with actionable insights for infrastructure planning, renewable integration, and decarbonization strategies, advancing methodological frameworks for sustainable energy transitions in Asia.
In this study, osseointegration has emerged as a revolution in the rehabilitation of prosthetic implants; it offers direct skin anchorage of the implant to the remaining limb and removes the limitations of conventional socket-based systems. The method improves the load transfer, proprioception, and comfort. FEM has been used for the validation process. These findings showed that the SACH foot exerted the largest vertical impact forces at first contact, resulting in high interface pressure at the bone-implant junction. The single-axis foot was better than the SACH in step stability and walking speed. A biomechanical comparison was conducted among three prosthetic foot types (SACH, single-axis, and multi-axis) within the framework of an osseointegrated transfemoral prosthesis. The case involved a 32-year-old male patient weighing 78 kg, who underwent a right above-knee amputation. Ground Reaction Force (GRF) tests were performed to determine the distribution of forces during walking, and bone-implant interface pressure measurements were performed to assess load transfer efficiency. The multi-axis foot, in contrast, exhibited the most even GRF distribution, with much lower interface pressure, indicating greater shock absorption and better functional comfort in transfemoral amputees. From the numerical analysis, Von Mises stresses were found to be 57.858 MPa for the SACH foot, 49.593 MPa for the single-axis foot, and 45.460 MPa for the multi-axis foot. The safety factors associated with 4.3, 5.0, and 5,4, respectively, affirm that all the designs do not exceed safe biomechanical ranges. Regarding total deformation, the multi-axis foot exhibited the lowest value (0.1837 mm). These numerical results are in harmony with the experimental results. The SACH foot transmits higher loads but also causes greater stresses and deformations within the implant system, whereas the multi-axis foot offers better mechanical safety, structural stability, and more uniform stress distribution. Overall, the findings indicate that the SACH foot has greater push-off power. Conversely, the multi-axis foot does not impose additional load on implants and increases safety over the long term, making it the most preferred choice among transfemoral amputees with Osseointegrated prostheses.
Life Cycle Assessment (LCA) is a widely recognized and essential methodology for evaluating the environmental performance of materials and processes throughout their life cycle, from raw material extraction, manufacturing, and use, to end-of-life treatment or disposal. In this study, LCA is systematically integrated into the development and performance evaluation of high-surface-area activated carbon derived from sugarcane bagasse via a dry chemical activation, with a specific focus on its application in Adsorbed Natural Gas (ANG) storage systems. The bagasse-based activated carbon exhibited a high specific surface area of 1,576 m2/g and an average pore diameter of 2.37 nm. To enhance methane storage capacity and desorption efficiency, the carbon was further modified using silver (BAC-Ag3), nickel oxide (BAC-Ni3), and barium oxide (BAC-Ba3). Among these, BAC-Ba3 demonstrated the most favorable performance, adsorbing 0.34 g/g and desorbing 0.31 g/g of CH4 at 26 bar and 27 °C. The Freundlich isotherm model projected a maximum adsorption capacity of 0.51 g/g at 35 bar. A cradle-to-gate LCA was conducted for the BAC-Ba3 synthesis process to quantify environmental trade-offs and identify impact hotspots. Results revealed that the most significant environmental burden stemmed from the Photochemical Ozone Creation Potential (POCP), with a value of 8.46 × 10-4 kg ethene equivalent. This study highlights the importance of integrating advanced material performance with comprehensive environmental impact assessments to promote informed decision-making for the sustainable design of adsorbents in ANG systems and to contribute to the broader development of clean energy storage technologies.
Accurate solar radiation predictions are a fundamental prerequisite for optimizing the performance and operational stability of photovoltaic power grids, particularly in tropical regions like Bali, which possess high solar potential but face significant challenges due to cloud-induced intermittency. This study aims to develop a high-precision forecasting model for hourly solar radiation in North Bali using a Deep Learning approach based on a Bidirectional Long Short-Term Memory (BiLSTM) architecture. Regarding materials, this research uses a comprehensive hourly climatological dataset sourced from the NASA POWER project, spanning January 2020 to January 2025. The dataset incorporates six critical meteorological features, including Air Temperature (T2M), Relative Humidity (RH2M), Solar Zenith Angle (SZA), and the Clearness Index, to capture atmospheric dynamics. The research methodology is systematic, encompassing data pre-processing, feature correlation analysis, and a rigorous model training phase. A key contribution of this work is the implementation of automated hyperparameter optimization using the Keras Tuner library to determine the optimal configuration of neuronal units and learning rates. The results demonstrate that the optimized model achieves superior predictive accuracy with a Root Mean Squared Error (RMSE) of 22.12 W/m², outperforming the baseline configuration which recorded an RMSE of 22.38 W/m². These findings confirm that a multivariate deep learning framework, when properly tuned, can effectively resolve complex nonlinear solar patterns. Consequently, this robust model offers significant practical implications for enhancing grid management strategies and renewable energy integration in Indonesia, while future work is suggested to explore hybrid Transformer-based architectures.
Improving the quality of oil palm seedlings is crucial because it impacts the overall sustainability and production of the crop. This study aimed to determine the impact of Trichoderma asperellum SL2 on the growth of oil palm seedlings in nurseries. The study consisted of a single factor: the doses of T. asperellum SL2 in mycelial biomass. This treatment factor consisted of four levels, namely without T. asperellum SL2, doses of 100 g/L, 200 g/L, and 300 g/L. There were three replications per treatment, resulting in 12 experimental units. The total plant population was 48 seedlings. Morphological and physiological traits were analyzed statistically using ANOVA. When the DMRT was administered after the treatment, it had a notable impact at p < 0.05. The measurements were taken on fully mature leaf samples taken with the LI-COR 6400XT from the fifth branch at the growth point. The optimal seed height of 32.06 cm was achieved with the 300 g/L Trichoderma asperellum SL2 treatment. Similarly, other observed variables, such as stem diameter (14.33 mm), the number of fronds (5.33 units), the longest root (34.83 cm), and root volume (9.83 cm³), were also best in the treatment with 300 g/L T. asperellum SL2. The 300 g/L T. asperellum SL2 produced a photosynthetic rate of 30.94 μmol CO2 m-2 s-1. Stomatal conductance was measured at 56.82 mmol H2O m-2s-1, and WUE at 24.69%. These findings collectively emphasize the promising prospects of applying T. asperellum SL2 treatment to elevate overall growth performance.
The Enhanced Diagnostic Visualization (EDV) system is an advanced medical imaging platform designed to improve neuroimaging analysis through efficient image loading, automated contrast enhancement, and interactive visualization tools. The proposed Graphical User Interface (GUI) is developed specifically to enhance the visualization of brain lesion images. It supports batch loading of Digital Imaging and Communications in Medicine (DICOM) files, enabling seamless management of large datasets from various imaging modalities such as MRI and CT scans. A feature named Dynamic Scan Layer Navigation (DSLN) Mechanism is introduced to facilitate loading and navigation through a series of brain scans, providing a detailed examination of brain structures across different axial planes and aiding the accurate identification of neurological conditions. A key feature of the EDV system is its automated contrast enhancement and visualization engine, which applies multiple histogram equalization techniques, including Dualistic Sub-Image Histogram Equalization (DSIHE), Exponential Logarithmic Histogram Equalization (ELEHE), Exponential Logarithmic Adaptive Histogram Equalization (ELEAHE), and Contrast Limited Adaptive Histogram Equalization (CLAHE). Additionally, colorization methods are also included to enhance grayscale medical images, improving the differentiation of tissue structures and anomalies. By generating precomputed multi-method contrast views, the system enables healthcare professionals to efficiently compare different enhancement techniques, streamlining the diagnostic workflow. The potential of this GUI lies in its ability to merge user-friendliness with powerful image processing capabilities, offering a comprehensive and efficient solution for medical professionals in clinical and research settings.
The network, computer, and Internet Network Security (IOT). Many challenges because of the evaluation of attack methods. The intrusion detection system (IDS) is considered. An important part of data protection, however, intrusion detection systems are compromised due to imbalanced data sets, high false alarms, and poor accuracy when using multiple large datasets. To address these contestations, this study suggests a new model based on the Jaccard-Based Pattern Similarity Approach. The similarity measures this model relies on are based on dividing data into small blocks to check similarities between feature sets and behavioral patterns in network traffic, achieving high detection accuracy and reducing false alarms. Experiments were conducted using the UNSW-NB15, Telemetry Data TON-IoT, and Wireless Sensor Networks WSN-DS datasets. Using common metrics (Accuracy, Precision, Recall, F1-score). The model performed excellently with an accuracy hitting 99.78% with the dataset (UNSW-NB15), accuracy for dataset (TON-IoT), 99.36% and accuracy for dataset (WSN-DS)99.07%. The proposed model's superiority in developing high-performance detection systems makes it a suitable option for cybersecurity. The results demonstrate that this model performs very well across diverse, ever-changing network environments. They also highlight how the approach significantly enhances intrusion detection, especially in IoT-based systems and wireless transmission networks.
Effective ecotourism destination management requires a dynamic understanding of visitor preferences. However, there is often a significant gap between quantitative satisfaction scores and the nuanced sentiments expressed in qualitative feedback. This study aims to develop an integrated smart evaluation system to bridge this gap, transforming raw data into evidence-based recommendations. The system is built on a validated assessment instrument with 21 items and three main factors: Value of Natural Beauty (VNB), Socio-Economic and Cultural Sustainability (SECS), and Educational and Environmental Awareness (EEA). Applying the Knowledge Discovery in Databases (KDD) framework, the system uses the K-Means clustering algorithm, with the optimal number of clusters (k = 3) determined by the Elbow Method, for tourist segmentation. Sentiment analysis of text feedback is performed using an advanced Transformer-based NLP model (IndoBERT) for contextual accuracy. The analysis successfully identified three significantly different tourist segments: Hard Ecotourists, Structured Ecotourists, and Soft Ecotourists. A key finding of this study was the revelation of a “satisfied-critical” phenomenon in the largest market segment (Structured Ecotourists). Although this segment provided high quantitative scores, IndoBERT sentiment analysis revealed a highly dominant negative sentiment (71%) focused on operational and facility issues, indicating a discrepancy between measurable satisfaction and perceived experiences. The result demonstrates that integrating quantitative analysis with context-aware NLP yields meaningful insights for managerial decision-making. The proposed system delivers personalized strategic recommendations for each segment, enabling managers to adopt a more adaptive, proactive, and data-driven approach to destination management.
The paper introduces a unified category of high-order exponential-time difference (ETD) numerical methods, combined with rational approximations of trigonometric functions, to efficiently and accurately solve nonlinear wave equations arising in electrical engineering and applied physics. The equations of nonlinear waves are used to model a wide variety of physical processes, such as the propagation of electromagnetic waves, plasma oscillations, optical fiber communication, and the transmission of signals in dispersive media. The suggested framework combines diagonal Padé rational approximations of the exponential of matrices with the required trigonometric correction terms, in particular to improve accuracy and stability in oscillatory wave phenomena. This combination of the precise trigonometric structure of wave propagators with the rational approximation of the same has produced ETD-Padé-Trig (ETD-PT) schemes that exhibit zero phase error on the imaginary axis, machine-independent energy conservation, and optimal order convergence. A stringent stability study proves that the proposed schemes are A-stable, and convergence estimates establish stiff-order optimality with uniform error constants, regardless of the spatial stiffness parameter. Three nonlinear wave equations are discussed in detail: the nonlinear Schrodinger equation (NLS), which is an equation governing optical pulse dynamics, the Klein-Gordon equation, which is an equation describing interactions of relativistic fields, and the Korteweg-de Vries (KdV) equation, an equation describing the dynamics of long-wave solitons. Extensive empirical studies carried out using MATLAB uphold fourth-order convergence in time, energy conservation to within 10⁻¹⁰ error in the long-term integration of the equations, and the realism of soliton interactions and Peregrine breather rogue-wave physics. The superiority of the suggested scheme in phase accuracy, energy preservation, and computational efficiency is demonstrated by comparisons with the classical Cox-Matthews ETD4 scheme and explicit Runge-Kutta methods when long-time wave simulations are required.
This review offers a critical review of current artificial intelligence (AI) and natural language processing (NLP) techniques that are increasingly applied to the humanities and social analysis, with a focus on low-resource languages, Arabic data, and culturally contextualized interpretations. This research synthesizes foundational AI/NLP research to define method families and survey an applied corpus of peer-reviewed humanities articles that apply AI/NLP to study war, exile, heritage, media, and identity. Across major task families (sentiment analysis, topic modeling, discourse and stance analysis, diachronic semantics, machine translation), the study finds common concerns with validity, including lack of transparency in datasets; insufficient details on data preprocessing; poor evaluation designs; and poor reproducibility practices. The report is evaluated using a quality appraisal. It is applied in direct AI/NLP applications that show interpretive strengths. However, there is a lack of data documentation, justification of baselines, and code or model sharing. The article concludes with reporting recommendations and a checklist to help authors and reviewers publish and review transparent, replicable, context-sensitive AI/NLP research in the humanities and social sciences.
Coffee is one of the most widely consumed beverages globally, and it is increasingly recognized for its potential as a functional food due to its rich content of bioactive compounds. Coffee contains polyphenols, diterpenes, and alkaloids that have been linked to numerous health benefits, including antioxidant, anti-inflammatory, and neuroprotective properties. Understanding how different brewing methods influence the abundance of these compounds is critical for maximizing health benefits and aligning coffee consumption with personalized nutrition strategies. This systematic review aims to assess which brewing method, such as drip, French press, cold brew, or regular immersion, results in the highest concentration of caffeine and chlorogenic acids, which provide health benefits. This systematic review follows PRISMA 2020 guidelines. A search of the Scopus, PubMed, and Web of Science databases was conducted, and studies focusing on different Arabica coffee brewing methods and their effects on caffeine and chlorogenic acids were screened. The extracted data include roast level, brewing time, temperature, coffee-to-water ratio, and grind size. Risk of bias was assessed using Risk of Bias in Non-randomized Studies–of Interventions (ROBINS-I). Findings suggest that hot brewing consistently produced higher levels of caffeine and chlorogenic acid than cold brewing. This review has identified brewing methods and parameters that produced the highest caffeine and chlorogenic acids. This research contributes to the growing field of precision functional foods, in which dietary components are customized for individual health benefits, offering a novel approach to dietary personalization that extends beyond standard nutritional recommendations.
This study evaluates the thermal performance of a portable water filtration system that integrates solar heating with natural coagulation–flocculation using Moringa oleifera seed powder. Moringa seeds contain cationic proteins that bind fine suspended particles and lignin in highly turbid water, forming larger flocs that settle more easily. In this system, the coagulation effectiveness of Moringa is enhanced by elevated water temperature generated by a simple solar collector comprising black-painted aluminum plates, copper pipes, and a glass cover. Raw water with an initial turbidity above 500 NTU was heated for 60 minutes. The results show that increasing the temperature from 27°C to 87°C significantly accelerated coagulation, flocculation, and sedimentation. Higher temperatures reduced water viscosity, increased particle collision frequency, and enhanced floc formation, allowing the water to clarify more rapidly. The solar collector maintained a thermal efficiency of 88–90%, supporting consistent heating throughout the process. Combined thermal and filtration treatment reduced Escherichia coli levels from 400 CFU/100 mL to 0 CFU/100 mL within 60 minutes. Turbidity also decreased markedly from over 500 NTU to below 30 NTU after final filtration, meeting the Indonesian Drinking Water Standards under Ministry of Health Regulation No. 2 of 2023. These findings demonstrate that solar-assisted heating greatly enhances the performance of Moringa as a natural coagulant. The developed system effectively improves water clarity and microbiological safety without relying on electricity, making it a practical and sustainable solution for clean-water provision in remote and off-grid communities.
This research emphasizes that the fundamental load transfer entities, specifically the connections, constitute the most critical details for designers handling advanced metal structures in their core production lines. This is primarily because stress concentration phenomena typically present a significantly more complex challenge concerning the safety of steel structures. The thermal (3D) numerical model employed for the 4φ and the particle count derived from this model demonstrate high accuracy. Results reveal that, for each compass point, the maximum peak von Mises stress under axial compression loads varies by approximately 35% relative to the combined loading condition. Additionally, localized stresses are predominantly concentrated around bolt holes and plate edges. Furthermore, an extensive parametric analysis indicates that an increase in maximum deformation from 10 mm to 20 mm, coupled with single-sided load eccentricity, results in an additional stress accumulation of about 28% compared to the maximum deformation alone. These findings underpin principles essential to steel connection design and performance, extending beyond conventional practices prescribed by codes. Load eccentricity, in particular, contributes to heightened stress concentrations by generating more localized stresses near the connection region. It is generally crucial for steel plate connection designs to account for various loading scenarios. The proposed finite element model provides valuable insights into stress distribution and structural response—details often absent in traditional analytical approaches. Such insights are instrumental for engineers aiming to optimize connection design, considering aspects such as geometry, plate thickness, and external load alignment. Ultimately, these advancements can enhance the safety and durability of steel structures over time.
