The Vitrified Clay Pipe (VCP) is a ceramic material that has been widely used in the last century to build sanitary sewer pipelines. The high sintering temperature in Vitrified Clay Pipes (VCPs) has resulted in high energy consumption during the VCP manufacturing process. The mechanical strength, which is influenced by the sintering process, is also a crucial factor in VCP performance. This study investigated the properties of clay and clay reinforced with Iron (III) Oxide (Clay/Fe2O3) at various weight percentages from 5 wt.% to 25 wt.%. The properties of clay and Clay/Fe2O3 composites were also investigated at different sintering temperatures from 950 degrees C to 1120 degrees C. The maximum shrinkage was 7.93% at a 75/25 clay/Fe2O3 ratio sintered at 1120 degrees C. The Modulus of Rupture (MOR) was also increased with sintering temperature. The maximum MOR of 45.54 MPa was exhibited at 80/20 Clay/Fe2O3 ratio sintered at 1120 degrees C. The shrinkage and the MOR results were in line with the microstructure obtained through Scanning Electron Microscopy (SEM) analysis. X-ray diffraction (XRD) analysis indicated that the main components of Clay/Fe2O3 composite were quartz, muscovite, and hematite phases. The study's findings suggest that Iron (III) Oxide (Fe2O3) has the potential to significantly improve the properties of clay for VCP application at lower sintering temperatures, thereby reducing energy consumption in the VCP manufacturing industry.
The increasing adoption of asynchronous online learning has intensified the need for intelligent and responsive learning support within learning management systems (LMS). However, most existing LMS platforms remain limited to content management and administrative functions, providing minimal real-time and contextual assistance for learners. This study presents the design, integration, and technical evaluation of an artificial intelligence (AI)-based online assistant integrated directly into an LMS to support asynchronous learning activities. The proposed system, named ANSIA, was developed using a systems engineering approach that enables seamless integration without modifying the core LMS architecture. ANSIA utilizes learning context data from the LMS to generate relevant and context-aware responses to learners' queries. The system was implemented as a web-based module and evaluated through technical performance testing focusing on response accuracy, response time, and system reliability under realistic asynchronous learning scenarios. The results demonstrate that the AI assistant achieves high response accuracy, acceptable response times under varying loads, and stable operation during continuous use. These findings indicate that integrating an AI assistant into an LMS can enhance learning support while maintaining system stability and scalability. This study contributes a practical and technically validated approach for developing adaptive, modular, and sustainable AI-assisted online learning systems.
This research endeavours to construct a blockchain-based carbon record-keeping system wherein non-fungible tokens (NFTs) are employed as digital assets to elevate transparency and accountability in carbon emissions documentation within the palm oil plantation industry. Leveraging the Solana blockchain, the framework converts carbon datasets into immutable NFTs, thereby preserving the authenticity and provenance of the information. Methodological procedures comprise a comprehensive literature survey, rigorous system analysis, detailed design, iterative implementation, and a multi-layered testing regimen encompassing black box, application programming interface (API), and token validation assessments. Outcomes indicate the framework's proficiency in reliably logging carbon metrics, facilitating the NFT minting procedure, anchoring carbon datasets to both the blockchain and the interplanetary file system (IPFS), and interoperating effortlessly with the Phantom Wallet and analogous digital wallets. Furthermore, the architecture maintains efficient component intercommunication and ensures the cryptographic integrity of the recorded carbon information. Collectively, the proposed system is anticipated to substantially streamline the logging and auditing of carbon emissions within the plantation sector, while concurrently equipping stakeholders to navigate the evolving carbon trading regulatory landscape in Indonesia.
A hybrid framework for accurate remaining useful life (RUL) prediction of lithium-ion batteries is developed by integrating Akima-based empirical mode decomposition (Akima-EMD) with a CNN-Bayesian LSTM model. Existing approaches, such as EMD-LSTM and EMD-CNN-LSTM, have improved prediction accuracy but remain constrained by limitations of traditional EMD, including overshoot, envelope distortion, and sensitivity to irregular degradation patterns. To address these issues, Akima-EMD is employed as a preprocessing method to decompose and denoise capacity signals while preserving meaningful degradation trends. A one-dimensional convolutional neural network (1D-CNN) is used for feature extraction, and a bidirectional LSTM is adopted to capture temporal dependencies. In addition, Bayesian optimization is applied to automatically tune key hyperparameters, improving model robustness under noisy and limited data conditions. The proposed framework is validated using NASA Ames Prognostics Centre of Excellence datasets (B0005 and B0006) under varying training conditions. Experimental results demonstrate that the proposed model consistently outperforms conventional LSTM and CNN-LSTM models, achieving the lowest end-of-life prediction error when combined with Akima-EMD preprocessing. These findings indicate that the proposed framework effectively handles nonlinear and irregular degradation patterns, providing a robust and reliable solution for battery RUL prediction and contributing to predictive maintenance in energy storage and electric vehicle systems.
Concrete production relies heavily on Ordinary Portland Cement (OPC), which depletes natural resources and contributes to environmental challenges. To address this, the present study explores the use of Rice Husk Ash (RHA) as a supplementary cementitious material and polypropylene (PP) fibres as reinforcement to improve concrete performance. The research problem centres on reducing cement consumption while overcoming the brittle nature of concrete. The objective is to evaluate how different lengths of PP fibres affect the mechanical properties of RHA-modified concrete, with the guiding question: how fibre length influences compressive and tensile strength in concrete containing RHA as a partial cement replacement. M15 grade concrete was prepared using a mix ratio of 1:2:4 with a water cement ratio of 0.50, where cement was replaced with RHA at 5%, 10%, and 15%, and 0.1% PP fibres of 0.5", 1.0", and 1.5" were added in mixes with 10% RHA. A total of 84 cube specimens were tested for workability, compressive strength, and splitting tensile strength at 7 and 28 days. Results showed that workability decreased with increasing RHA and fibre length, compressive strength improved with 10% RHA (6.4% higher than control) and was further enhanced by PP fibres, with 0.5" fibres giving the highest gain (17.8%). Splitting tensile strength also improved with RHA and PP fibres, with 0.5" fibres achieving the maximum increase (46.7% over control).
This study developed and evaluated an innovative Online Aptitude Test and Mock Board Examination system designed to address critical academic integrity challenges prevalent in high-stakes online assessments for nursing students. Following a Design Science Research paradigm and employing Agile Software Development, the platform evolved from a web application to a native mobile application, which embeds the existing WebContent within a secure container. It integrates robust cheating prevention features like dynamic randomization and obfuscation of questions, mandatory audio and video recording using RecordRTC, and OS-level screen capture prevention. Automated post-exam proctoring is performed by a server-side Python analysis pipeline. This pipeline leverages FFmpeg to remux and index recordings, applies OpenCV and dlib for frame-by-frame face and head-pose detection, and uses the audioop module to calculate silence and noise ratios, outputting structured JSON reports for anomaly flagging. Pilot implementation demonstrated the system's efficiency, achieving an 86.05% audio/video transmission success rate and which, through its analysis, identified three confirmed cheating incidents among 355 examinees. Independent evaluations against ISO/IEC 25010 Software Product Quality Standards consistently yielded "Excellent" quality ratings from both system users with an average weighted mean of 3.68 and IT experts with an average weighted mean of 3.66, affirming its functional suitability, usability, and performance. The results conclusively establish the developed system as a secure, scalable, and adaptable digital assessment solution, effectively upholding academic integrity while simultaneously enhancing examinee board performance.
This study develops a mobile-based learning media to support supervision in physical education and sports training. The development followed seven stages: research design, user needs analysis, application design, development, trial and evaluation, revision, and implementation/dissemination. User requirements were collected from supervisors, coaches, and teachers to determine core functions for supervision, including structured learning modules, activity documentation, evaluation quizzes, scoring logic, and progress monitoring. The application was implemented and piloted in a PJOK teacher working group context, using a pretest-posttest scheme and a usability evaluation based on the System Usability Scale (SUS). The results indicate increased user knowledge after using the application and a mean SUS score of 77, corresponding to a good usability level. The developed media provides practical support for supervision workflows by enabling access to learning materials, standardized evaluation, and progress tracking through a mobile interface. These findings suggest that mobile learning media can improve the efficiency and consistency of supervision practices in physical education and sports training.
NEBULA is a web-based electronic learning medium designed to support personal energy-balance awareness by integrating energy intake and expenditure to promote sustainable, healthy lifestyles among university students. This study examined the effectiveness of Design Thinking-STEM (DT-STEM) learning supported by NEBULA in enhancing students' problem-solving skills. A quasi-experimental pretest-posttest control group design was implemented by comparing an experimental group receiving DT-STEM with NEBULA and a control group receiving conventional STEM learning. Problem-solving skills were assessed using essay-based instruments aligned with established problem-solving indicators. The findings indicate that students in the experimental group demonstrated higher problem-solving performance across multiple aspects, particularly in visualizing problems, describing problems in scientific terms, and executing solution plans. These results suggest that technology-assisted DT-STEM learning through NEBULA effectively fosters problem-solving competence and 21st-century skills, while simultaneously supporting Sustainable Development Goals (SDGs).
The spread of Internet of Things (IoT) has revolutionized many fields by enabling extensive connection options and data -driven decisions. IoT network expansion had introduced significant security weaknesses, which required strong security mechanisms. Integration of artificial intelligence (AI) into IoT, called artificial intelligence of things, provides increased ability to detect and attenuate the intelligent danger. This study presents an ensemble-based Intrusion Detection System (IDS) utilizing a soft voice technique to improve the detection of the cyber attackers in the IoT network. CIC-IOT2023 and IOTID20 data sets were evaluated by empirical evaluation, including ransomware and various types of cyber-attack. The proposed ensemble model achieved better performance with the accuracy rate of 95.06% and 99,998% respectively on CIC-IOT 2023 and IOTID20 data sets, with high precision, recall, F1-score and ROC-AUC values indicating better performance. Comparative analysis demonstrated that the proposed model improves individual models and improves the reliability of safety systems and efficiency. These findings suggest that ensemble learning represents a promising opportunity to detect future infiltration in the smart environment.
X (formerly Twitter) remains a central yet polarized platform for online interaction, particularly within Indonesia's growing digital landscape. To assess public perception, this study analysed 2,794 cleaned Indonesian reviews from the Google Play Store. Following lexicon-based labelling, data were vectorized using term frequency-inverse document frequency (TF-IDF) and Bag-of-Words. Logistic Regression and Random Forest classifiers were evaluated (80:20 split), utilizing the Synthetic Minority Over-sampling Technique (SMOTE) to mitigate a 24.39% class imbalance. Logistic Regression with TF-IDF and SMOTE achieved optimal performance, yielding 90.2% accuracy and a 92.2% F1-score (90.8% precision, 93.8% recall), significantly outperforming Random Forest (83.0% accuracy). The high recall specifically indicates effectiveness in correctly identifying relevant sentiment. Consequently, this study demonstrates that lightweight linear models combined with over-sampling provide a robust, computationally efficient framework for analysing Indonesian text, enabling developers to derive actionable insights for strategic product improvement and user retention.
The growing need for sustainable energy in the oil and gas industry has led to increased interest in renewable power sources. Vortex-Induced Vibration (VIV) energy harvesting offers a promising method for converting kinetic energy from fluid flow into electricity. The research focuses on qualifying the influence of structural stiffness on the performance of a single rigid circular cylinder in a VIV energy generation system. Numerical simulations were performed for a Single Degree of Freedom (SDOF) model using Computational Fluid Dynamics (CFD). The Reynold number of 37000 and stiffness values ranged from 60 N/m to 800 N/m were considered. Results indicate that an optimal stiffness of 600 N/m yields a peak power output of 6.10 W, demonstrating that structural stiffness is a critical factor in maximizing energy conversion efficiency. These findings offer valuable guidance for designing VIV energy generation system to capture energy from water flow more efficiently for clean and sustainable energy.
This study investigates how corrosion affects the buckling response of AA7075 aluminium columns under increasing dynamic axial loads. Both long and intermediate columns were examined to assess the role of slenderness ratio in such conditions. Experimental work involved testing multiple specimens under dynamic compression, including standardized buckling tests, with comparisons made between corroded and intact samples. Lateral deformation was measured accurately using a digital dial gauge placed at 0.7 of the column length from the fixed end. The analysis relied on established models, including Euler's equation for long columns, Perry-Robertson for intermediate columns, and the Euler-Johnson approach for transitional ranges. According to design standards, lateral deflection was limited to 1% of the effective length. To replicate corrosion effects, specimens were buried for 60 days before testing. The findings indicate that corrosion reduces structural performance. After 60 days of exposure, the ultimate compressive strength decreased by 1.78% compared to uncorroded specimens. The critical buckling load also declined with increasing corrosion time, reaching a maximum reduction of 4.24%. Comparison with theoretical models showed that the Perry-Robertson equation, using a safety factor of 1.3, best matched the experimental data, while the Euler-Johnson and Euler equations were less accurate for intermediate-short and long columns, respectively. Finite element analysis in ANSYS closely matched experiments, particularly with a 1.1 safety factor. Results confirm corrosion weakens aluminium columns, while reliable models predict buckling behaviour. Additionally, AkzoNobel Intergard epoxy effectively protects AA7075 columns, preserving performance in saline environments.
The increasing demand for compact, low-loss, and high-performance microwave filters in modern wireless systems presents challenges in achieving miniaturization without compromising signal quality. This work proposed the design and analysis of a Substrate Integrated Waveguide (SIW) bandpass filter utilizing a circular cavity structure. The objective is to optimize the filter's performance in terms of bandwidth, centre frequency, return loss (S11), and insertion loss (S12). A key focus is placed on the TE21 resonant mode, which significantly influences the electromagnetic response of the circular cavity. The resonant behaviour is governed by critical geometric parameters, including the cavity radius, via diameter, via spacing, and feedline dimensions. The filter is designed to operate at 2.4 GHz; a frequency commonly used in wireless communication systems. A Taconic CER-10(TM )substrate, measuring 0.64 mm in thickness and possessing a dielectric constant of 10, is utilised to attain compactness and superior performance. Electromagnetic modelling findings indicate areturn loss (S11) of-18.02 dB and an insertion loss (S12)of-1.38 dB at the designated frequency, signifying effective impedance matching and little signal attenuation. Compared to conventional rectangular SIW filters, the proposed circular cavity configuration achieves improved field confinement and reduced footprint, highlighting its novelty and practical advantage. To validate the design, fabricated prototype measurements are compared with the simulated results. The comparison confirms the accuracy of the simulation model and highlights the potential of circular SIW cavity structures in developing efficient and compact microwave filters for modern communication systems.
Sundanese script is a type of script that is increasingly rarely used, and therefore, many people are no longer familiar with it. The difficulty in recognising Sundanese script arises due to variations in handwriting forms. The aim of this study is to measure the performance of densenet-121 in recognizing Sundanese handwriting, both in the form of letters and combinations between basic letters and rarangk & eacute;n (XXX). The approach adopted in this study involves using the DenseNet-121 model to classify 17,280 handwritten Sundanese script images into 144 classes of script images, after applying grayscale conversion, thresholding, segmentation, and resizing. This model was evaluated based on accuracy, precision, recall, and F1-score. The test results showed that the best model utilised the AdamW optimiser, a learning rate of 0.001, and 10 epochs, which produced an accuracy of 79.5%, a precision of 63.2%, a recall of 65.3%, and an F1-score of 62.0%. The evaluation results showed fairly good accuracy. However, there was a caveat: the model suffers from bias. This was due to the high diversity of the data. Unique features in the data need to be generalized. Compared to previous studies, the larger dataset in this study provided broader coverage, but also added complexity that needs to be considered to improve model performance in the future.
This study presents a comparative evaluation of calcium copper titanate (CCTO)-polydimethylsiloxane (PDMS) composites fabricated via solvent and nonsolvent techniques, focusing on their morphological, chemical, and dielectric properties relevant to high-frequency microsystem applications. Composites containing 10-60 wt.% CCTO were prepared using ethanol-assisted dispersion (solvent method) and direct blending (nonsolvent method). Morphological and elemental analyses via scanning electron microscopy (SEM) and energy-dispersive X-ray spectroscopy (EDX) revealed improved filler packing and reduced aggregation in nonsolvent-processed samples. Fourier-transform infrared spectroscopy (FTIR) confirmed physical filler-matrix interactions without evidence of chemical bonding. Broadband dielectric measurements across 1-10 GHz showed that nonsolvent composites consistently exhibited higher permittivity and lower dielectric losses. At 60 wt.% loading, the nonsolvent sample demonstrated a dielectric constant of 5.49 and a loss tangent of 0.0417 at 5.2 GHz, alongside an electrical conductivity of 0.066 S/m. These results suggest that solvent-free fabrication enhances dielectric performance while simplifying processing, providing a practical approach for designing flexible substrates for RF and microsystem integration.
Energy model calibration is essential in assessing building energy performance and developing effective energy management strategies. To enhance calibration scenarios worldwide, the adaptability of building components, such as indoor air temperature (IAT) and heating, ventilation, and air conditioning (HVAC) systems, must be evaluated across diverse building types, including educational buildings. However, this aspect remains underrepresented in existing studies. Evidence suggests that combining simulated and measured data improves calibration accuracy in building energy simulations, ultimately leading to optimized energy management and thermal comfort. Given the above-mentioned information, this study introduces a methodological approach to energy model calibration for educational buildings. The main objectives are to establish a reliable baseline model, to design a robust IAT measurement framework, and to integrate measured and simulated data for model calibration. The proposed approach incorporates architectural drawings, building envelope properties, and comprehensive data collection as key inputs for developing a reliable energy model baseline. The methodology also includes a detailed data measurement framework that specifies the duration, instruments, and sensor placements to ensure robust results. EnergyPlus simulation software was used to create the baseline model, integrating standardized parameters, such as thermostat setpoints. Furthermore, two sets of IAT measurements were collected using data loggers placed in the southern and northern zones of the building, respectively. The findings reveal a strong correlation between Sensor-S1 and Sensor-S2 measurements, thereby validating their reliability for monitoring IAT. Baseline calibration under stable conditions also demonstrated the model's effectiveness, while measurements during transient events highlighted the importance of realtime monitoring, sensor placement, and adaptive HVAC strategies. The study's findings underscore the importance of incorporating dynamic behaviours and occupancy data to improve accuracy in educational buildings. This calibrated energy model also serves as a valuable tool for benchmarking and predicting energy performance, thus offering a foundation for future research and enhanced energy management strategies across diverse building types.
Demand management in information technology (ITDM) complements existing software development methods, facilitates resource allocation, and requires an organisation with clearly defined roles and skills. However, efficient ITDM requires an understanding of the factors that are essential to its success. Consequently, the purpose of this article is to determine the critical success factors (CSF) that influence ITDM. This research study follows a quantitative, non-experimental design. Data were collected from 144 public institution employees through a self-administrated survey using a questionnaire, and a Likert scale was used to measure the factors associated with ITDM. Structural equation modelling (SEM) was used for this analysis, which revealed that ITDM is influenced by CSF such as top management support, strategic alignment between IT and business, IT portfolio alignment, and leadership. This study provides valuable insights for professionals and decision-makers on the CSF influencing ITDM, highlighting the importance of their prioritisation. It contributes to the existing literature on ITDM by providing a deeper understanding of the importance of these CSF in the ITDM process.
The identification of plant-based biostimulants requires integrative approaches that link community-derived knowledge with experimental validation. This study integrates social media knowledge mining, high-performance liquid chromatography (HPLC) phytohormone profiling, and germination bioassays to identify plant-derived biostimulants. Indonesian local plants were screened from TikTok, Instagram, and YouTube based on growth-related claims, then processed into aqueous filtrates. Endogenous indole-3-acetic acid (IAA), gibberellic acid (GA3), and trans-zeatin were quantified using HPLC, and biological efficacy was evaluated through mung bean (Vigna radiata) germination assays at different filtrate concentrations. The results revealed pronounced interspecific variation in phytohormone profiles and significant improvements in germination percentage, uniformity, seedling growth rate, and vigor index for several filtrates, particularly at 50-75% concentrations. Multivariate analyses demonstrated clear associations between hormonal composition and germination performance. This integrated framework provides a robust and scalable approach for identifying plant-based biostimulants by bridging digital knowledge mining with analytical and biological validation.
Cryptographic S-boxes are essential components of cryptographic primitives that induce confusion by making the relationship between the secret key and the ciphertext complex and obscure. An S-box must satisfy the Strict Avalanche Criterion (SAC), meaning that flipping a single input bit should alter each output bit with a probability of 0.5, ensuring that the output appears random and uncorrelated with the input. However, designing an S-box to meet this ideal is challenging, and the difficulty increases when it must also be efficiently masked using Threshold Implementation (TI). In this paper, we identify S-boxes that not only exhibit perfect SAC but also maintain uniformity when masked with TI. By constraining the algebraic degree of the S-boxes to 2, we achieve uniform TI implementations through degree-count classification of TI shares. Furthermore, our threshold implementation method applies to other degree-2 5-bit S-boxes while preserving uniformity. From the generated S-boxes, we select an S-box namely P , that scores exceptionally well in nonlinearity, SAC, BIC, differential uniformity and linear probability. Our results contribute to the field of cryptographic security by providing a practical solution for constructing S-boxes that meet SAC requirements and integrate seamlessly with TI, thereby enhancing the security and effectiveness of side-channel countermeasures. Our work provides new insights for studies that aim to achieve uniformity in TI implementations of algorithms whose algebraic degree exceeds 2.
Proppant performance plays a decisive role in sustaining fracture conductivity during hydraulic fracturing. Conventional sand, while inexpensive, exhibits poor sphericity, low mechanical strength, and severe fines generation, whereas ceramic and conventional resin-coated proppants, though more durable, remain costly and less widely applicable. This laboratory-scale study introduces apolyurethane-based nanocoating reinforced with carbon nanotubes (PU-CNT) as a scalable alternative to overcome these limitations. In this laboratory-scale study, sand proppants were coated with a polyurethane-carbon nanotube (PU-CNT) nanocomposite prepared via probe sonication and applied using a spray-coating technique. The nanocoated sand proppants exhibited notable morphological improvements, with sphericity and roundness increasing from 0.47 to 0.73 and 0.13 to 0.30, respectively. Surface wettability shifted from strongly hydrophilic 12 degrees to hydrophobic 92 degrees, enhancing hydrocarbon mobility. Mechanical testing confirmed superior crush resistance with fines generation reduced to 0.34%, significantly outperforming uncoated and conventional resin-coated counterparts. These results highlight the originality and practical significance of PU-CNTs coated proppants, offering a balance of costeffectiveness, durability, and performance compared with current industry options.
