Dergi Yayın Koleksiyonu

Bu koleksiyon için kalıcı URI

https://hdl.handle.net/20.500.12395/13187

Listeleniyor 1 - 20 / 440

Water content classification on cucurbitaceae family fruits from VIS/NIR spectroscopic data
(Selçuk Üniversitesi, 2024) Ürgen, Nurullah; Demircan, Semiye
Visible and Near-Infrared (Vis-NIR) spectroscopy is a technique used to determine the chemical and physical properties of matter by analyzing electromagnetic radiation across a broad wavelength range, specifically from 400 to 2500 nm. In this application, the aim is to assess the quality attributes of six Cucurbitaceae family fruits, namely: zucchini, bitter melon, ridge gourd, melon, chayote, and cucumber, using a single classification model for all fruits rather than individual models. This classification model predicts whether it exceeds 90% according to fruits based on water content. Samples with water content above 90% are labeled as high-water content, while those below are categorized as low-water content. For preprocessing, Standard Normal Variate (SNV) and Neighborhood Components Analysis (NCA) methods were employed to optimize the feature space. The model was trained using a Support Vector Machine (SVM) classifier. Without feature extraction, the accuracy ranged from 90% to 92.5%; however, with feature extraction, the accuracy increased to 95%-97.5%. This classification model successfully predicts high water content, an essential indicator of product quality and productivity, across the dataset with high precision. By integrating comprehensive data processing and machine learning techniques, this study demonstrates a reliable method for assessing product quality, contributing significantly to the field of agricultural and food industry quality control.
The modal and dynamic analysis of sprayer booms with diverse structural geometries
(Selçuk Üniversitesi, 2024) Buğdaycı, Ali Rıza; Şen, Muhammed Arif
In spraying agricultural areas with a sprayer boom, the vibrations on the boom due to irregularities and variable speeds in the agricultural land directly affect spraying performance. Vibrations on the system disrupt the spraying pattern and cause agricultural yield to decrease. In addition, it may be subject to deformations due to structural strains caused by vibration, which may cause performance losses and long-term failures. In this study, the modal analysis of a sprayer boom with the same working width but a different geometric structure, which is widely used in agricultural spraying, was performed with the finite element method its natural frequencies and mode shapes were obtained, the dynamic responses of the system under harmonic strains were examined, and the obtained simulation results were evaluated comparatively. Firstly, the solid model of the system was obtained, transferred to the finite element program, and its modal properties were simulated. Afterwards, the sprayer was fixed to the body connection and forced with a harmonic acceleration input determined by considering the irregularities in the fields. The obtained results are presented in detail with graphs. From the simulation results, it is understood that the structural geometry affects the mode shapes and natural frequencies of the system. It is thought that the results can be useful in the structural optimization of the system, in the control of vibration dynamics and can contribute to the development of new design strategies.
Leveraging predictive analytics for operational efficiency in automotive after-sales services
(Selçuk Üniversitesi, 2024) İkizler, Tuğçe; Özçelik, Abdullah Engin; Uslu, Banu Çalış
This research explores the application of predictive analytics to optimize operational efficiency in the automotive after-sales sector, focusing on inventory management and workforce allocation. By employing ARIMA and SARIMA models, seasonal and trend-based forecasts were generated using data collected from multi-brand service centers between 2018 and 2021. The results demonstrated a strong seasonal influence on service demand, with peaks identified in the second and fourth quarters, aligning with routine maintenance patterns. Key findings revealed a 32% dependency between technician numbers and spare part usage, while daily replacement volumes ranged from 89 to 327 parts, requiring precise workforce planning during peak periods. The originality of this research lies in its integration of predictive analytics into after-sales service management, an area where empirical studies are scarce. Unlike traditional approaches, this study not only highlights the significance of after-sales services in customer satisfaction but also provides actionable insights for cost reduction and resource optimization. For instance, the forecasting models facilitated dynamic inventory management, reducing holding costs while maintaining service reliability. Additionally, seasonality analysis guided the efficient allocation of technicians, minimizing operational downtime and improving customer experiences. These findings underscore the transformative potential of predictive analytics in the automotive industry. By leveraging data-driven insights, businesses can enhance their operational resilience and competitiveness, laying the groundwork for more sustainable and efficient service systems. This research addresses a critical gap in the literature by demonstrating how predictive models can directly contribute to strategic decision making in after-sales services.
Experimental evaluation of thermal conductivity and thermogravimetric analysis of Jatropha Oil-based titanium nano-cutting fluid
(Selçuk Üniversitesi, 2024) Nwachukwu, Victor C.; Lawal, Sunday Albert; Saka, Abdulkareem Ambali; Okoro, Uzoma Gregory
Nanoparticles have several potential applications due to their advantageous properties which have been identified as the main driving force for nanofluids research. In this study, Jatropha oil, extracted from the seeds of the Jatropha plant was characterized by investigating the physico-chemical properties. The Jatropha oil was thereafter used as the base fluid for nanofluid formulation by enhancing with Titanium oxide (TiO2) at 0.1%, 0.15% and 0.2% volume concentrations. The formulated nanofluid was characterized by evaluating the thermal conductivity and degradation profile (thermogravimetric analysis). The findings revealed that the locally sourced Jatropha oil has 0.916 Specific gravity, 7.85 mg/100g Acid value, 189.33 mgKOH/g saponification value, 2190C flash point, -70C pour point, 5.09 pH, 113.4 g/100g of KOH iodine value and 32 mm2/s viscosity at room temperature. It was also found that the nanoparticle cutting fluid enhanced with TiO2 as a better thermal conductivity at 0.15% concentration compared to the pure base fluid and other enhanced nanofluid modified with 0.1 and 0.2% TiO2 concentration. In addition, the thermogravimetric analysis results (TGA and DTG) revealed that the pure jatropha oil degraded fastest with a broad peak and a wider degradation temperature range (226.12-449.69°C) compared with modified nanofluid with a smaller degradation temperature range (229.11-438.33). Therefore, it was concluded that the nanoparticle cutting fluid modified with TiO2 (0.15% concentration) can be adopted as cutting fluid for machining operations.
Skin lesion segmentation with semantic SAM: Pros and cons
(Selçuk Üniversitesi, 2024) Gül, Sevda; Aydın, Bekir Murat; Akgün, Devrim; Kara, Rabia Öztaş; Çetinel, Gökçen
The Segment Anything Model (SAM), introduced in April 2023, has gained prominence for its ability to generalize across various image segmentation tasks. This study evaluates SAM's performance on skin lesion segmentation using both public (3463 images) and private (773 images) dermoscopy image datasets, the latter collected with ethical approval from *** University Training and Research Hospital. The segmentation performance was assessed using Intersection over Union (IoU) and Dice metrics, achieving Dice scores of 0.6598 (IoU: 0.5865) for the private database and 0.6513 (IoU: 0.5624) for the public database. A post-processing step was applied to refine the segmentation results, enhancing SAM's ability to delineate lesion boundaries. However, while SAM demonstrated strong generalization, its performance on low-contrast and irregularly shaped lesions indicates the need for further adaptation. This paper highlights SAM’s potential in medical image segmentation while outlining its limitations, especially in specialized tasks like skin lesion analysis.
Internet of things (IOT) enabled drip irrigation system (DIS) for the growth of Allium Fistulosum
(Selçuk Üniversitesi, 2024) Atojunere, Eganoosi Esme; Omotoro, Boluwatife
The project was designed to develop an automated Drip Irrigation System(DIS) for the application of water for the growth of Allium Fistulosum. Allium Fistulosum is known for its economic values as a medicinal plant however reaching its daily water requirement is becoming difficult especially due to insufficient amount of rainfall and also shortage of manpower when water is applied manually. Allium Fistulosum’s seed was planted on a 5 m by 5 m land plot beside the Department of Systems Engineering, University of Lagos, Nigeria in May 2023.Data of rainfall pattern, evapotranspiration of Lagos, Nigeria were collected from metrological agency to calculate the Evapotranspiration(ETo) and Reference Crop Evapotranspiration(ETcrop) for Allium Fistulosum to define the threshold level for the DIS to irrigate. The moisture meter probe inserted into the soil layer continuously monitored the moisture content of soil and interfaced reading with the M-32 microcontroller that periodically uploaded to the cloud (ThingSpeak) for visualization. The DIS for irrigation was fabricated at the Control Laboratory consisting moisture meter, MSP-32 microcontroller, relay, DC submersible pump, water reservoir all connected together to computer system. Results showed that ETo and ETcrop for Allium Fistulosum was 5.4mm/day and 4.86 mm/day at crop coefficient (Kc) of 0.9 which was the threshold set for Pump to irrigate by releasing the required amount of water to the Allium Fistulosum. If the moisture content was above the threshold, no action is required from the pump. The efficiency of the DIS was found at 65 to 80%, worked according to the specifications and was user friendly. The DIS can be scaled up to accommodate larger farm plots to be irrigated which would reduce human involvement in irrigation.
Modeling of dynamic testing of steering system tie rod assembly using hydraulic system
(Selçuk Üniversitesi, 2024) Baysal, Selim Sefa; Akdemir, Bayram
Steering system components transmit motion from the steering wheel to the wheels. Vehicle rotations consist of moving parts made up of ball joints. Ball joints operate at specific angles and torques. These parts are exposed to forces from the road and the vehicle during operation. Articulated parts must be tested dynamically. The backlash and torque values measured before the test must remain within the appropriate limits after the applied life test. Tie rod end and axial joint piece tested at the same time. The tie rod end and the axial joint were connected to each other and tested in the test setup. Purpose Before and after the test, it is to determine how many millimeters of movement in the body and what the torque change is for the articulated parts in both parts. Before the test, the breaking torque of the tie rod end was 7.75 nm and the operating torque was 3.12 nm. Tie rod end clearance was measured between 0.007 millimeters in the axial positive direction and -0.006 millimeters in the negative direction in the static test. The breaking torque of the pretest axial joint was 11 nm and the operating torque was 4 nm. In addition, the gap value was determined as 0.039 millimeters in the positive direction and -0.071 millimeters in the negative direction. In the fatigue test, the change in torque and clearance values for both parts remained below 10 percent at a frequency of 5 Hz and at the end of one million cycles, and the accuracy of the production parameters was ensured. The control of the hydraulic test system was checked with the Labview program and reporting was made after the test. In the study, the rod parts working on the vehicle were simulated successfully.
Friction Stir Welding (FSW) and Friction Stir Spot Welding (FSSW) on polycarbonates: A review
(Selçuk Üniversitesi, 2024) Aslan, İbrahim; Yaka, Harun
In recent years, it is seen that the use of polymers has become widespread in the field of manufacturing. This means that polymers are as important as metals in the manufacturing sector. Polycarbonates, which are thermoplastic polymers, are widely used, especially in automotive, construction and health fields. Today, it is well known that polycarbonate can be bonded in a variety of ways, including bonding and welding. In this study, an investigation was made on the application of polycarbonate polymers in friction stir welding (FSW) and friction stir spot welding (FSSW), which can properly join the same or different types of materials and have economic advantages. In this paper, the studies in the literature on the joining of polycarbonate materials with each other or with other materials by FSW and FSSW are reviewed and discussed by specifying the production parameters and details used. Thus, a single point of reference is provided for researchers working in the field. In this paper, useful findings have been shared and useful for researchers to get an idea about the joinability of polycarbonate polymers with FSW and FSSW with different materials that have not been tried before.
Combining LSTM-enhanced features with machine learning algorithms for improved heart failure prediction
(Selçuk Üniversitesi, 2024) Acar, Züleyha Yılmaz; Tok, Ümit
It is well-known that the majority of deaths in the world are caused by heart disease. Therefore, early diagnosis of heart disease is of vital importance. Artificial intelligence techniques that aim to support specialists are among the most effective methods used in the field of health. In this study, in order to improve the detection of heart failure, we proposed a classification scheme to improve heart failure detection by generating new representations of the dataset using the LSTM model (LSTM-enhanced features) and machine learning algorithms (support vector machine (SVM), k-nearest neighbor (kNN), naive bayes (NB)). The LSTM was used to extract deep features that reveal the dependencies among the dataset. The 11 features from 918 data samples in the dataset were re-represented with LSTM and used as 100 LSTM-enhanced features. Experimental results showed that our proposed scheme achieved an accuracy of 92.90%, precision of 94.90%, recall of 92.08%, and F1-score of 93.47%. Performance comparisons with other studies demonstrated that the LSTM-based scheme proposed in this study is applicable to similar datasets.
An advanced online appointment system for hairdressers and barbers to communicate seamlessly with their customers
(Selçuk Üniversitesi, 2024) Avuçlu, Emre; Taptal, Kevser
One of the biggest problems faced by busy modern people today is not being able to keep up with the pace of time. It can sometimes be difficult to make an appointment during the day to go to the hairdresser and barber, which is one of our basic needs. Potential customers of hairdressers and barbers may not be able to come to the workplace during working hours. They may want to make an appointment in their own free time and at convenient times. In this study, an advanced online appointment system was developed for hairdressers and barbers. With the developed application, customers can easily make appointments and cancel their appointments when necessary. In addition, a penalty system is applied to customers for canceled and missed appointments according to the determined rules. Customers in this situation cannot make new appointments for a certain period of time. Additionally, the application has dynamic functions such as searching for hairdressers, managing employees and setting working hours. With the appointment management developed in this application, the workload of hairdressers and barbers is alleviated and customer satisfaction is increased. Thanks to the detailed management features in the system, business owners can track their employees and customers more effectively. The online appointment system contributes to the digitalization process by replacing traditional methods and thus increases the competitiveness of businesses.
An advanced dynamic web-based automation application for instructors and students to track lessons and homework
(Selçuk Üniversitesi, 2024) Avuçlu, Emre; Sarıoğlan, İsa
Web-based teaching system is a type of distance education and is a teaching system in which internet technologies are used to transfer course material to the student. It is known that it is difficult to keep track of homework assignments and course files at every level of educational institutions, especially at education levels where class sizes are large. This situation causes instructors to spend extra time. Due to limited time, the homework control carried out by the instructors cannot be carried out completely or the homework control is incomplete. Since instructors mostly carry out homework control in the classroom, the feedback given by students to homework is insufficient. In this study, a web-based application was developed to facilitate the exchange and communication of documents in courses where instructors give homework and projects. With the developed web-based application, transactions such as sending files digitally, tracking homework, and grading can be done between teachers and learners. Thanks to the web-based homework-project tracking system, instructors will spend less time in the homework and project evaluation process. Thus, they will save time for other academic or other educational work. Additionally, unnecessary paper usage will be prevented.
Feature extraction and recognition on traffic sign images
(Selçuk Üniversitesi, 2020) Çınar, İlkay; Taşpınar, Yavuz Selim; Sarıtaş, Mücahid Mustafa; Köklü, Murat
It is vital that the traffic signs used to ensure the order of the traffic are perceived by the drivers. Traffic signs have international standards that allow the driver to learn about the road and the environment while driving. Traffic sign recognition systems have recently started to be used in vehicles in order to improve traffic safety. Machine learning methods are used in the field of image recognition. Deep learning methods increase the classification success by extracting the hidden and interesting features in the image. Images contain many features and this situation can affect success in classification problems. It can also reveal the need for high-capacity hardware. In order to solve these problems, convolutional neural networks can be used to extract meaningful features from the image. In this study, we created a dataset containing 1500 images of 14 different traffic signs that are frequently used on Turkey highways. The features of the images in this dataset were extracted using convolutional neural networks from deep learning architectures. The 1000 features obtained were classified using the Random Forest method from machine learning algorithms. 93.7% success was achieved as a result of this classification process.
Automatic brain tumor segmentation with k-means, fuzzy c-means, self-organizing map and otsu methods
(Selçuk Üniversitesi, 2020) Aşlıyan, Rıfat; Atbakan, İsmail
The human brain is an amazing organ of the human nervous system and controls all functions of our body. Brain tumors emerge from a mass of abnormal cells in the brain, and catching tumors early often allows for more treatment options. For diagnosing brain tumors, it has been benefited mostly from magnetic resonance images. In this study, we have developed the segmentation systems using the methods as K-Means, Fuzzy C-Means, Self-Organizing Map, Otsu, and the hybrid method of them, and evaluated the methods according to their success rates of segmentation. The developed systems, which take the brain image of MRI as input, perform skull stripping, preprocessing, and segmentation is performed using the clustering algorithms as K-Means, Fuzzy C-Means, Self-Organizing Map and Otsu Methods. Before preprocessing, the skull region is removed from the images in the MRI brain image data set. In preprocessing, the quality of the brain images is enhanced and the noise of the images is removed by some various filtering and morphological techniques. Finally, with the clustering and thresholding techniques, the tumor area of the brain is detected, and then the systems of the segmentation have been evaluated and compared with each other according to accuracy, true positive rate, and true negative rate.
Big, medium and little (BML) scheduling in fog environment
(Selçuk Üniversitesi, 2020) Bichi, Bashir Yusuf; Islam, Saif Ui; Kademi, Anas Muazu
Fog computing has got great attntion due to its importance especially in Internet of Things (IoT) environment where computation at the edge of the network is most desired. Due to the geographical proximity of resources, Fog computing exhibits lower latency compared to cloud; however, inefﬁcient resource allocation in Fog environment can result in higher delays and degraded performance. Hence, efﬁcient resource scheduling in Fog computing is crucial to get true beneﬁts of the cloud like services at the proximity of data generation sources. In this paper, a Big-Medium-Little (BML) scheduling technique is proposed to efﬁciently allocate Fog and Cloud resources to the incoming IoT jobs. Moreover, cooperative and non-cooperative Fog computing environments are also explored. Additionally, a thorough comparative study of existing scheduling techniques in Fog-cloud environment is also presented. The technique is rigorously evaluated and shows promising results in terms of makespan, energy consumption, latecny and throughput.
Development of an agriculture robot for row-type seed sowing applications
(Selçuk Üniversitesi, 2020) Yurtsever, Cihan; Ertaş, Yasin; Sustam, Oben; Ulu, Cenk
In this study, the design and development of an agriculture robot which has row type seed sowing feature are presented. The robot consists of four subsystems; a four-wheel mobile platform, a digger mechanism, a seed dropping mechanism, and an irrigation mechanism. The electrical and mechanical designs of the robot are performed depending on the specified design criteria. System control software and user interface are developed considering stakeholder expectations. Designed subsystems are manufactured and integrated. Furthermore, robot functionality tests are performed and the desired performance of the agriculture robot is validated by the test results. The robot is remotely operated via an Android application on a mobile phone and all operation data can be monitored via this android application. Additionally, the robot can perform the row-type seed sowing operation in an autonomous mode. The developed agriculture robot has the potential to provide an efficient and inexpensive way for future seed sowing applications.
Vertex cover based link monitoring techniques for wireless sensor networks
(Selçuk Üniversitesi, 2020) Dağdeviren, Züleyha Akusta
Wireless sensor networks (WSNs) are generally composed of numerous battery-powered tiny nodes that can sense from the environment and send this data through wireless communication. WSNs have wide range of application areas such as military surveillance, healthcare, miner safety, and outer space exploration. Inherent security weaknesses of wireless communication may prone WSNs to various attacks such as eavesdropping, jamming and spoofing. This situation attracts researchers to study countermeasures for detection and prevention of these attacks. Graph theory provides a very useful theoretical basis for solving WSN problems related to communication and security issues. One of the important graph theoretic structures is vertex cover (VC) in which a set of nodes are selected to cover the edges of the graph where each edge is incident to at least one node in VC set. Finding VC set having the minimum cardinality for a given graph is an NP-hard problem. In this paper, we describe VC algorithms aiming link monitoring where nodes in VC are configured as secure points. We investigate variants of VC problems such as weight and capacity constrained versions on different graph types to meet the energy-efficiency and load-balancing requirements of WSNs. Moreover, we present clustering and backbone formation operations as alternative applications of different VC infrastructures. For each VC sub-problem, we propose greedy heuristic based algorithms.
Thruster design for unmanned underwater vehicles
(Selçuk Üniversitesi, 2020) Gücer, Çetin Arda; Acar, Onur; Kantarcıoğlu, Burak; Ulu, Cenk
Underwater researches have been carried out for various purposes such as the protection and investigation of natural and environmental resources, various construction activities, finding and extracting fossil fuel resources, academic and industrial researches. Especially in the last two decades, unmanned underwater vehicles are effectively used in almost all of these researches. One of the most essential parts of those vehicles is their thrust system which gives them the ability to move underwater. In this study, the design of a thruster for unmanned underwater vehicles is given. The designed thruster system consists of four main parts: an electric motor, a driver circuit, a magnetic coupling transmission element, and a propeller. The electrical and mechanical designs of these parts are performed depending on the predetermined design criteria. A brushless type DC motor is chosen as an electric motor, and the required torque and rpm values are determined analytically. Depending on the chosen electric motor, a suitable driver circuit is determined. Then the propeller, the magnetic coupling element, and the motor housing are designed by using the SolidWorks software package. Pressure and fluid dynamics analyses of the housing and propeller are performed by using the Ansys software package. The thruster design is validated by simulation results.
Design and development of a test setup for reaction wheel systems of nanosatellites
(Selçuk Üniversitesi, 2020) Koç, Muhammed Hayri; Ünlü, Umut; Kuvat, Tarık; Ulu, Cenk
Nanosatellites have gained an important place in space applications thanks to developing technology. For a successful operation, attitude determination and control systems in satellites are vital. A reaction wheel system is the widely used drive system for nanosatellites. An electric motor driven reaction wheel is a system that operates utilizing from conservation of momentum and law of action and reaction. In this study, the design and development of a test setup for reaction wheel systems of nanosatellites are given. By using this test setup, different configurations of reaction wheels can be tested, performances of different control methods can be evaluated, and the energy efficiency of the whole system can be determined. Additionally, measured test data such as orientation angles and system current, voltage, and power can be recorded and monitored via the developed user interface. The test setup consists of a platform, reaction wheels, and a control unit. The mechanical design of the test setup which allows changing reaction wheel configurations is developed in Solidworks software. Modeling and control studies are performed in Matlab Simulink environment for brushless dc motor driven reaction wheels. The electronic control unit is designed, and Raspberry Pi is used as a controller. The test platform is produced by using 3d printer and then, subcomponents (electrical control equipment) are assembled into the platform. The functionality and performance tests of the system are performed successfully. The PD control performance results for attitude control of the satellite with the specific reaction wheel configuration are given. These results match the simulation results and validate the system design.
Next generation display technology: Transparent organic light emitting diodes
(Selçuk Üniversitesi, 2020) Albaba, Mohammed; Akkuş, Meryem Sena
Transparent displays have attracted significant interest as new generation display technology in diverse fields, such as glasses, automotive industries, or military technology applications. These applications need simultaneous projection or display of data and visibility of the surroundings through the device to reach transparent OLED, the colour of the emission layer materials, and the opaqueness of the metal thin film cathodes are the important aspects that have to be addressed. In the last few years, Transparent Organic Light Emitting Diode (TOLED) display got very big attention in terms of developing and improving its transparency to be able to make the background visible to the user as much as possible. In this paper, a review of some studies in the literature about Transparent Organic Light Emitting Diode (TOLED) has been made. Some information about prototype applications and technology has been given. Some factors are affecting by making the vision clear in the TOLED such as; haze, the distance between the TOLED display and the object behind it, illumination, transparency, and contrast. There are also some studies about using TOLED displays in vehicles. It recommended that transparency should be ideal and clear as well for the driver.
Dimension and color classification of olive fruit with image processing techniques
(Selçuk Üniversitesi, 2020) İnce, Fatma Betül Kınacı; Taşdemir, Şakir; Özkan, İlker Ali
The development of image processing technology appears in agriculture as well as in many other fields. Various classifications are carried out for fruits and vegetables. These are processes such as determining the harvest time according to their degree of maturity, deciding the way of collection and performing packaging operations according to their dimension. This study aims to classify the fruit according to its intended use in order to benefit more from the olive fruit that is important in industrial terms. In this study, olive fruit is classified as big, medium, and small according to its dimensions. Also classified as black and green according to their colors. This classification process was made in MATLAB environment and the KNN algorithm and decision trees was used. The results are obtained with Euclid and Manhattan methods used with the KNN algorithm and are given comparatively. According to the application results, 100% success was achieved in both methods in color classification. In dimension classification, 89.2% classification success was achieved in KNN algorithm and 86.7% in decision tree method.

Güncel Gönderiler