Polyurethane (PUR) foam layers with polyvinylchloride (PVC) coating used in bus instrument panels are subjected to various mechanical and thermal effects. Understanding the behavior of these materials under these effects is critical to improve product performance and durability, and for this purpose, finite element analysis (FEA) was employed to complement experimental testing. The aim of this study is to investigate the mechanical and thermal behavior of PVC-coated PUR specimens by FEA and in particular to investigate the degradation mechanisms and the performance of the material under different conditions. The mechanical behavior of PVC coated PUR specimens was modeled under different loading and temperature conditions and analyzed with finite element software. The simulations were modeled based on experimental data and literature information and included tensile, three-point bending, peel and thermal expansion tests. Strain-displacement ratios were monitored using digital image correlation (DIC) techniques. The results are in high agreement with the experimental data. Strain quantities show small variations due to calculation errors, but 97% convergence in three-point bending tests and 96% convergence in single lap joint tests. Thermal analysis and DIC results confirm the other findings. The work presented here demonstrates the ability to simulate the behavior of PVC-coated PUR specimens through FEA, avoiding material and time waste and contributing to design changes. FEA results were in good agreement with the physical tests, confirming suitability for specimen-level analysis.

Truck-mounted mobile cranes are specially designed systems used for lifting and transporting loads of various weights and dimensions. The physical dimensions of these systems may vary depending on the areas of use and different technical requirements. This variation can significantly affect both the lifting capacity and the operating speed of the mechanism. Dimensional changes may lead to undesirable deviations within the safety limits defined during the design phase. Therefore, optimizing the dimensions of these mechanisms requires careful evaluation of the statistical nature of manufacturing and measurement errors. In this study, in order to achieve dimensional optimization of crane boom lifting mechanisms, the Monte Carlo simulation method was combined with the Particle Swarm Optimization technique, and the optimum dimensions of the mechanism were determined based on reliability.

This study addresses the hospital site selection problem -strategically vital for the effective delivery of healthcare services-through a microbiological risk-based multi-criteria decision-making (MCDM) approach. The research focuses on the Halfeti district of Þanlýurfa, Türkiye, evaluating five candidate locations based on six comprehensive criteria: microbiological risk, population density and demand, land cost, accessibility, infrastructure adequacy, and environmental/disaster risk. Criterion weights were objectively determined using the Fuzzy Entropy method, while the ranking of alternatives was conducted using the Interval Type-2 Fuzzy TOPSIS method to better capture uncertainty and the variability of expert opinions. The analysis revealed that the Halfeti district center (A1) is the most suitable location for a hospital when all criteria are considered. Among the criteria, microbiological risk (weight: 0.268) and environmental/disaster risk (weight: 0.415) emerged as the most influential, indicating the critical role of health and safety concerns in site prioritization. In contrast, social criteria such as population density and accessibility received relatively lower weights, reflecting the specific demographic and infrastructural characteristics of the region. The findings offer a robust and comprehensive evaluation framework under uncertainty, which can be applied to similar healthcare investment decisions in rural or semi-urban settings. Moreover, the study highlights the broader applicability of Interval Type-2 fuzzy logic-based models in areas such as disaster management, environmental planning, and public infrastructure investments.

Multi-Criteria Decision Making (MCDM) methods provide systematic approaches for evaluating alternatives under multiple criteria. Determining the relative importance of criteria is a critical step that directly affects the reliability of the obtained results. In this study, the Extended Standard Deviation (ESD) method is proposed to overcome the limitations of the classical Standard Deviation (SD) method. The proposed method offers a more comprehensive weighting process by considering not only the internal variations of individual criteria but also their interrelationships with other criteria. Unlike conventional SD, the ESD method calculates weights based on both the individual distributions of criteria and their effects on other criteria. This approach enables a more holistic evaluation of the degree of contrast among criteria and the overall structure of the dataset. The primary objective of this study is to conduct a comparative analysis of the proposed method against the classical SD method and other widely used objective weighting techniques, thereby identifying their respective advantages and limitations. To assess the applicability of the proposed method, sensitivity, comparative, and simulation analyses were performed, and the method was statistically evaluated by applying it to different decision matrices. The findings indicate that the proposed method provides a robust and reliable alternative in objective weighting processes.

In this study, the Family Travelling Salesman Problem is considered and time constraints are included in the model to better represent real-life applications. The mathematical model for the proposed problem has been adjusted as necessary and a metaheuristic method has been developed in order to achieve good solutions in shorter times. The method is a Variable Neighbour Descent algorithm using four different neighbourhood structures and a tabu list is added to the algorithm to be used in some neighbourhood movements to make the solution space search more efficient. The perturbation operator also diversifies the search by making large changes on the solution. The proposed algorithm was compared with the mathematical model results and performed better on the sample sets used.

This study deals with the evaluation of the load frequency control (LFC) by Sliding Mode Control (SMC) having two different structures. As the LFC is one of the important problems of the power systems, there have been many solutions proposed for this problem, especially based on the well-known PID controller. The SMC is an effective alternative that has been focused on for power systems. Therefore, two recently popular SMC algorithms are evaluated for the LFC of a single area power system. One algorithm is model-based, first-order SMC, and the other one is a model-free, second-order super-twisting SMC algorithm smoothed with a hyperbolic tangent function. Optimization of the controllers is performed with two metaheuristic algorithms, the Sine Cosine Optimization Algorithm and the Grey Wolf Optimizer. The controllers’ performance is evaluated for an applied 0.1pu load. Detailed results are given in tabulated form and graphically.

Image compression is a crucial technique for reducing storage requirements and improving transmission efficiency of digital images, especially given the ever-increasing volume of image data. However, conventional lossy compression methods such as JPEG and JPEG2000 often introduce significant quality degradation, particularly when compressing highly detailed images. This study presents an optimized wavelet transform-based image compression method designed to minimize information loss while maximizing compression efficiency. The proposed method integrates adaptive thresholding, the selection of optimized wavelet functions, and multi-level wavelet decomposition to address the limitations of traditional approaches. Specifically, adaptive thresholding is used to dynamically adjust compression parameters, reducing unnecessary data retention, while the wavelet function selection process ensures the most suitable basis for image features. Multi-level wavelet decomposition enables the retention of important image details across various resolution scales, improving compression without compromising visual quality. The performance of the proposed method is evaluated on several image types, including well-known test images, and compared against standard image compression techniques such as JPEG and JPEG2000. Experimental results show that the proposed method outperforms the conventional methods in terms of both compression ratio and image quality preservation, achieving higher Peak Signal-to-Noise Ratio (PSNR) and Structural Similarity Index (SSIM) scores. The proposed approach is particularly effective for applications requiring high-quality image storage and transmission, such as medical imaging, satellite imagery, and multimedia communication.

This study investigates the influence of rotor design on the performance of interior permanent magnet brushless DC (IPM-BLDC) motors used in axial fan applications, which span from industrial cooling systems to automotive technologies. Two alternative rotor topologies with a 12/8 slot–pole configuration were analyzed using finite element analysis (FEA). The evaluation focused on key performance metrics, including torque ripple, cogging torque, and efficiency. The optimized flux-barrier rotor demonstrated a 55.36% reduction in cogging torque, a 1.23% improvement in efficiency, and a 2.23% decrease in torque ripple relative to the baseline design. Based on these results, the superior rotor geometry was prototyped, and the numerical findings were experimentally validated through performance testing. The outcomes confirm that rotor flux-barrier optimization enhances the overall efficiency and operational stability of IPM-BLDC motors in axial fan applications.

Obtaining maximum efficiency from photovoltaic (PV) systems through maximum power point tracking (MPPT) remains an ongoing challenge. In this study, the weighted mean of vector (INFO) algorithm is employed to address and solve the MPPT problem for a photovoltaic system operating under partial shading. Besides INFO algorithm, electric eel optimization (EEFO), red-tailed hawk algorithm (RTHA), and student psychology-based optimization (SPBO) algorithms were also employed, and this study is the first to employ these optimization algorithms for MPPT purposes. The particle swarm optimization (PSO) algorithm, which is frequently employed in MPPT studies, is employed to compare the performance of new metaheuristic algorithms. These algorithms are tested with challenging shading scenarios where the local maximum points (LMPP) and global maximum power point (GMPP) varied. The performance of these algorithms is evaluated using the Friedman test, which is a statistical test, and performance metrics. According to the findings of the comparison, the INFO algorithm is the most effective among the five algorithms for MPPT optimization under partial shading conditions, and this conclusion is confirmed statistically. Additionally, experimental tests were conducted to evaluate the performance of the INFO algorithm on real hardware. A programmable PV simulator, boost converter, and STM32 board were used. The experiments demonstrated that the algorithm could quickly and stably track the maximum power point.

Two distinct Titanium Oxide based memristive devices were fabricated. One device was electrically characterized with a driving current, the other with a driving voltage. Two approaches were utilized for modelling these devices: firstly, novel models of I-V behavior were developed using curve fitting in MATLAB. Secondly, an existing memristor model, Quasi-Static Memdiode Model (QMM) was investigated and modified to reflect the behavior of the fabricated memristive devices. Thus, models for both current driven and voltage driven devices were extracted. SPICE and Verilog-A coding languages were used to simulate the devices in SPICE and Cadence Spectre to enable the simulation and design of hybrid memristor+ CMOS circuits on these widely used platforms. The accuracy of the models was verified by comparing simulation results to measurement results.

Sudden cardiac arrest (SCA) occurs when the heart becomes unable to pump blood effectively. Two of the most common arrhythmias that cause SCA are ventricular fibrillation and ventricular tachycardia. Treating these arrhythmias through defibrillation—delivering an electric shock to the heart—is vital for patient survival. Automated External Defibrillators (AEDs) analyze the patient’s heart rhythm and automatically deliver a shock when necessary. To do this, AEDs collect electrocardiogram (ECG) signals and use shock advisory algorithms (SAAs) to decide whether a shock is required. However, the development of SAAs suitable for AEDs involves challenges, such as the limited data processing capacity of embedded systems. The algorithm must be capable of reliably distinguishing heart rhythms in such constrained environments. A review of existing studies reveals both threshold-based and machine learning (ML)-based SAAs, with many ML-based algorithms demonstrating high classification performance. Yet, only a small fraction of these algorithms have been tested in real-time embedded systems, and their applicability to AED devices has been evaluated in limited contexts. In this study, a traditional ML-based SAA was developed using the random forest method and evaluated with a publicly available dataset. The proposed algorithm achieved 92.9% sensitivity for shockable rhythms and 99.2% specificity for non-shockable rhythms. Initially developed in a high-level programming language, the SAA was integrated into C and tested on a microcontroller-based development kit. With a memory requirement of 500 kB and a detection time of 75 microseconds, the algorithm was shown to be suitable for implementation in AED devices, demonstrating its potential for commercial use.

This study presents a question answering system developed for higher education using transformer-based models. Five pretrained models were evaluated including BERTurk Base cased/uncased, ELECTRA-Turk, mBERT and XLM-R. The models were fine-tuned on the THQuAD dataset and tested on a frequently asked questions dataset constructed from official university sources and student queries. In addition to standard evaluation metrics such as Exact Match and F1 score, an extended evaluation approach was applied to better capture semantically appropriate answers. ELECTRA-Turk achieved the highest F1 score of 0.8936 and an Exact Match score of 0.8478. The results show that transformer-based approaches can effectively support automated question answering in academic domains and improve information access for students.

Credit scoring prediction is critically important for financial institutions to effectively manage credit risk and ensure sustainable profitability. Accurate credit decisions require the development of predictive models based on historical data. In this study, predictive models were developed using various machine learning algorithms along with the Apriori algorithm based on association rule mining applied to a credit scoring dataset.
The modeling process leveraged data mining and artificial intelligence techniques; the performance of different classification algorithms was evaluated using 10-fold cross-validation through metrics such as accuracy, precision, recall, and F1-score. Statistical analyses (Wilcoxon signed-rank test and paired t-test) revealed that the Deep Neural Network (DNN) model outperformed Logistic Regression, Naive Bayes, and Random Forest models significantly, while exhibiting similar performance to MLP, SVM, and XGBoost models. This finding supports the strength of DNN models, particularly on complex datasets.
The results demonstrate the effectiveness of data-driven predictive approaches in credit risk analysis and emphasize the importance of selecting algorithms appropriate to the scenario. Furthermore, the advantages and limitations of the applied algorithms were evaluated, concluding that choosing the most suitable method according to the application context is critical. These findings contribute to credit risk prediction modeling efforts and provide guidance for financial institutions in developing AI-based decision support systems.

Hospital admission necessity of a patient who is under care for the possibility of arbovirus infection is a critical decision for healthcare practitioners. Medical staff may experience stress when making this decision due to the potential risks it poses to the broader community. Current capacities for diagnosis can be confusing. For this reason, data mining approaches have been proven to be highly effective in the diagnosis of diseases as well as in many other fields. As many research studies suggest, they can also be used to decide whether a patient with arbovirus infection should be hospitalized or not. For this purpose, this study uses Severity Index for Suspected Arbovirus (SISA) dataset and implements various machine learning classification techniques with the aim of binary classification to detect the hospitalization status of a specific patient. Several neural networks, single classifiers, and ensemble supervised learning methods are selected as classifiers during the experiments. The best classification accuracy value is obtained by Random Forest (RF) model with 0.9908. This model has been shown to outperform many data mining techniques previously applied in prominent studies. This improved result leads to additional experiments with a different number of estimators when implementing RF. The outcome also improves the maximum classification performance up to 0.9926 using 25 estimators. The study reveals the effectiveness of ensemble models, especially bagging and boosting approaches, for Arbovirus suspected case analysis.

Water demand and difficulty in accessing clean water resources increase the importance of effective and sustainable management of water losses in utilities. Regular measurement of data, analysis of effective factors and efficiency of methods and tools are quite important for sustainable management of losses. The aim of this study is to develop a current status assessment system for analyzing the data quality and monitoring the current implementation levels of apparent loss management components. The system consists of 45 components under main headings of the technical, operation and maintenance, commercial and economic based on literature and field experience. A scoring system is proposed to evaluate these components between 0 and 5 (quite good, good, insufficient, poor and quite poor). The system is tested with using field data in three utilities. The data quality and current status of apparent loss management practices in utility II (generally good) is better than other utility I (insufficient) and III (poor). The components that need improvement within the scope of apparent loss management are defined in utilities based the scoring results. This system provides the opportunity to question in detail the quality, measurement frequency and accuracy of the data used in the apparent loss management in utilities. However, the main problem experienced in the implementation of this system is the lack of sufficient report, data and records for apparent loss prevention, reduction and detection methods in utilities. This assessment system will help more realistic data for apparent loss management in the utility and to allow the field implementations more accurately.

Bomb attacks are occurring in the world due to the increasing hot and cold wars. In this study, the contact explosion, which affects the resistance of the structure the most, was investigated among these attacks. Since the explosive is in contact with the surface of the structure in contact explosion, the reaction of the structure in bending and shear behavior due to sudden dynamic loading differs from other loads. In our experimental study, 50x50x15 cm reinforced concrete slabs reinforced with steel wire mesh, wire fence and geogrid building materials were produced in order to compare the behavior of concrete against explosion. In contact with these plates, 577 gr explosive was applied and detonated. In the experiments, the contact explosion reactions of geogrid reinforced concrete and steel-reinforced concrete and unreinforced concrete were compared. As a result of the experiments, it was determined that the concrete reinforced with wire fence and geogrid is applicable against contact explosion.

In this study, the combustion characteristics, combustion index and activation energies were investigated by thermogravimetric analysis (TGA) of different proportions of fuel blends prepared using rose pulp, domestic sewage sludge and local Kale Lignite. In this study, three different fuel blends (15%C+15%C+15%G+70%L; 25%C+25%G+50%L; 40%C+40%G+20%L) were prepared. Thermal degradation of the fuel blends occurred in four regions. In this study, it was observed that the ignition temperatures increased with increasing sludge content in the blends, while there was no general change in the burnout combustion temperatures. The maximum mass loss rates (DTGmax) of the blends increased with increasing sludge content and provided a more stable combustion. In terms of combustion index, it was observed that the addition of sewage sludge to lignite improved the combustion characteristics of the fuel. The activation energies of the fuel blends prepared in this study were calculated as 86.68 kJ\/mol for 15%Ç+15%Ç+15%G+70%L, 107.20 kJ\/mol for 25%Ç+25%G+50%L, 114.07 kJ\/mol for 40%Ç+40%G+20%L and 160.05 kJ\/mol for 100% sludge. Therefore, it is seen that the thermochemical utilization of biomass and lignite with sewage sludge has a significant environmental and economic potential.

In our country, hard coal reserves used for coke production are quite limited. However, non-coking lignite reserves are quite high. In order for lignites to be added to hard coals and used in coke production, the coking properties of these blends should be determined. In this study, Erzurum Oltu lignite and Zonguldak Kozlu hard coal and blends of these coals prepared at certain ratios were used. Various parameters such as moisture content, ash, volatile matter, petrographic analysis, free swelling index, Gray-King coke test, dilatation and plasticity (fluidity) which determine the coking quality of raw coal and coal blends were investigated. The 75% huminite content of the maceral composition in the Oltu coal indicates its low-ranking character, while the 65% vitrinite content in the Kozlu coal indicates its high-ranking and coking character. Microlithotype compositions also support this situation. In the free swelling index test of the coals and blends, it was determined that 100% Oltu lignite did not show any adhesion and 100% Kozlu coal had a FSI value of 6, which reflects a good coking property. In the Gray-King coke test, Oltu lignite alone did not produce a coke type, while 100% Kozlu coal and 20% Oltu-80% Kozlu coal blend produced a G type coke. Fluidity tests revealed that Oltu lignite alone showed no plasticity or fluid behaviour, while Kozlu coal showed a wide plastic range of 87°C and a maximum fluidity of 980 DDPM (Dial Division per minute). In expansion tests, Oltu lignite showed softening and shrinkage behaviour without any expansion, while Kozlu coal showed softening and shrinkage as well as a significant expansion rate of +118%.