Theses supervised by Dr. Öğr. Üyesi Göktürk Poyrazoğlu

17 theses · Özyeğin University

Master'sOpen AccessEN

Applications of time-series methods in electricity price forecasting

The prediction of day-ahead electricity prices with higher accuracy is always helpful for the market players of the power exchange. Models were designed in the first place to find out the best time-series estimation method for the selected 12 European countries before and during COVID-19 using exponential smoothing family and Naïve processing. The classification approach is followed by 33 different features of each country to answer the question of the best method for other countries. Also, 24 hours are estimated without any independent variables using seasonal autoregressive integrated moving average. Apart from the statistical forecasting model, long-short term memory is used for each European country. The multiple linear regression is used on electricity price forecasting. Different predictors are analyzed to reduce the mean absolute percentage error. The training data includes the dates from the day-ahead electricity market in Turkey. It is proved that the lagged electricity prices such as previous one day, one week, and lagged moving average prices play a key role in electricity price estimation. Aside from other valuable coefficients, natural gas, oil, and coal prices are tested in the forecasting model. The error rates of the fuel prices are noticeably decreased by using multiple linear regression that generates more accurate results and crucial variables influencing hourly electricity price have been determined. Different training data length is an essential part of decreasing the error proportions in the electricity price estimation. Also, it is analyzed that there is no big difference regarding the error rates if it is compared to the Regular and Dynamic Regression models in the forecast of electricity prices. Regular and Dynamic Feedforward Neural Network Methods were utilized for the model to give better error performance and more accurate results were obtained.

Toygar Ülgen
Özyeğin University · Institute of Graduate Studies in Science
2021
00
Master'sOpen AccessEN

Yarım köprü llc rezonant çevirici topolojisi kullanarak OLED TV için ince güç kaynağı ünitesi tasarımı

Recently, with the development of technology, televisions which have with different structures and features have been developed in the field of TV. Tv technologies such as QLED, Miniled and OLED, which are developed as rivals to traditional LED TVs, continue to become widespread in the market everyday. OLED TVs do not have a backlight structure, and each pixel is radiated by itself, it enables a thinner design. Since each pixel is spontaneously luminous, it provides good picture quality and allows to see blacks more black and colors more clearly. OLED TV panels are generally supplied in the 20-28V voltage range. The higher power consumption is required with higher luminance and nits values increasing everyday. Power cards have certain mechanical limitations in design, such as thin TV design and limit of specific area. As a result, there is a need for a high efficiency card design that can operate reliably at high currents. The power supply is designed under three main blocks that is EMI filter and rectifier stage, interleaved PFC stage, half-bridge LLC coverter stage. Filter block structure is as C-L-C-L-C. An interleaved PFC block which has 390V output voltage is designed to transfer high power. For the PFC stage, the efficiency is aimed at %95.Temperature comparison has been made with different transformer types on the half-bridge LLC resonant converter. The advantages and disadvantages of using EFD and toroidal cores have been determined. EFD type core is an integrated transformer. When using toroidal transformer, extra series resonant coil is used to dissipate the heating. In this way, high power is transfered with toroidal transformer. Synchronous rectifier is used on the secondary side to increase efficiency . LLC resonant converter stage efficiency is aimed at %96. For the power supply, input voltage range is between 100VAC-270VAC and output power is 384W in 24V@16A.

Switched mode power supplyEnergy efficiencyPower converters+1
Samet Kurt
Özyeğin University · Institute of Graduate Studies in Science
2021
00
Master'sOpen AccessEN

Yarı rezonans devreli indüksiyonlu ocaklarda ısı transferini ve verimliliği artırmak

This thesis proposes a dual-mode coil structure for induction cookers using a single switching element quasi-resonant circuit and examines its effect on heating performance. In an induction cooker, the material base to be used for the heating process and the coil adaptation significantly affect the heating performance. Since the materials used are user-defined and vary, they may not always be suitable for traditional single-coil structures. The proposed dual coil structure works according to the base diameters of the material to be used and provides adaptation for different diameters. This adaptation optimizes the heat distribution on the coil, allowing the energy to be transferred more effectively. Effective energy transfer by minimizing losses improves the heating performance, making it faster and more efficient. At the same time, with this structure, the magnetic field created by the coil is applied to the material more effectively, reducing the voltage value affecting the switching element. The dual-mode coil structure with the quasi-resonant circuit was verified by simulating with MATLAB, and its improvement in heating performance in the induction cooker was experimentally observed.

Electromagnetic inductionEnergy efficiencyPower converters+1
Hüseyin Küçükosman
Özyeğin University · Institute of Graduate Studies in Science
2021
00
Master'sOpen AccessEN

Güç çeviricileri için kapasitif kuplajlı trafo tasarımı

Transformer types have varied over time according to the purpose of use. As it is known, these devices are coiled materials that provide power transfer by magnetic induction. The AC voltage applied to the primary is reflected on the secondary side, resulting in a voltage induced in the secondary by the magnetic field. These materials, which we encounter in many places in power systems and power electronics, are difficult to produce and take time, and are costly because a certain number of turns is required. In this study, instead of an inductive transformer model, a capacitive transformer model will be designed and tested. This device will transfer power with an electric field. With the mutual capacitance value created by the four metal plates being in the same dielectric material in a nested structure, the power transfer will be transferred from the primary plates to the secondary plates. It is expected that the capacitive transformer model here will be better than the inductive transformer in terms of manufacturing convenience and cost. This design will also provide isolation for some types of non-isolated converters. The equivalent circuit of this capacitive transformer was set up, a simulation study was performed with MATLAB Simulink, and a power transfer was performed as desired. Similarly, the experimental circuit was set up and the results were obtained. It has successfully done the purposes of step-down and isolation transformers.

Circuit simulationPower transformerPower converters
Necdet Kaan Onur
Özyeğin University · Institute of Graduate Studies in Science
2021
00
Master'sOpen AccessEN

Üç faz çiftyönlü elektrikli araç şarj devresinin modellenmesi ve simülasyonu

The aim of this research is to investigate how an electric vehicle (EV) charger with controllers performs and affects the EV battery, integrated to the grid. At present, typical applications of EV chargers usually include only charging of the vehicle battery providing unidirectional power flow between the power grid and EV. The proposed three-phase grid connected EV charger system enables bidirectional power flow, sending unused energy in EV battery back to the electricity grid in the scope of V2G operation in addition to conventional charging of the battery. The EV charger circuit consists of two stages, one of which is three-phase AC/DC PWM converter and the other is synchronous DC/DC converter, between which a DC link capacitor having 800V rated bus voltage is inserted. Also, a Li-ion battery pack with the nominal voltage of 500V is used for the system. LCL filter is utilized for AC/DC part of the charger to attenuate line current ripples and the other components are selected based on the theoretical analysis performed as well. Two-loop control algorithms, of which design parameters are calculated using some detailed mathematical analysis, are implemented for both AC/DC converter and DC/DC one to achieve tight voltage & current regulations for the EV charger. Then the performance of the bidirectional EV charger circuit is evaluated via some simulation work in Matlab/Simulink based on closed-loop operation and harmonic contents for grid side. Simulation results are discussed to analyze the closed-loop performance of the EV charger system during various conditions and the developed theory regarding control system design is verified. Harmonics & THD values of line currents are also investigated for various battery voltage levels using the obtained results. In addition, loss analysis is performed to show that the bidirectional EV charger system can operate with high efficiency values.

Görkem Berk Şahinler
Özyeğin University · Institute of Graduate Studies in Science
2021
00
Master'sOpen AccessEN

Mathematical modeling and control of asynchronous motor for multi-physical models

Asynchronous motor is a good choice for the industry due to its low cost, high efficiency, robustness, high reliability, ability to work in bad environmental conditions, easy recycling, good dynamic performance, high maximum speed, and high range of field weakening [1, 3]. Although the scalar control method is sufficient in speed control applications by keeping the stator voltage/frequency ratio constant in asynchronous motors, in cases where the continuity of torque control is important, the scalar control method is not sufficient and therefore cannot compete with the direct current method [4]. With the scalar control method, a large range of speed control could be made at the nominal torque of the induction motor. However, the scalar control method was not sufficient in applications where continuous speed and torque control were required. Current components forming the flux and torque in the direct current motor are controlled independently from each other. The excitation flux forming the flux can be kept constant and the moment can be controlled by the armature winding current. However, there are no such two perpendicular current components in an asynchronous motor. There is only stator current with amplitude angle and phase available. Therefore, the stator current is transformed into two perpendicular current components, one forming the flux and the other the torque, using Clarke and Park transformations. The desired torque control can be achieved by controlling these two current components separately. Closed-loop volt/hertz controllers were developed to control torque and speed. In this control method, the moment can be brought to the desired values in the continuous regime, but the moment in the transient regime cannot be controlled. The speed control method, which is sufficient in many applications, is not sufficient in applications where the torque should be controlled continuously. This study aims to perform mathematical modeling of asynchronous motor and simulation of the rotor flux-oriented vector control method on a user-friendly platform. Besides, the obtained model is created as a PhiSim library, a product of Sherpa Engineering. This library was created using the bond-graph theory. Asynchronous motor is a good choice for electric vehicles due to its simple structure, low cost, high efficiency, robustness, high reliability, ability to work in bad environmental conditions, good dynamic performance, and wide speed range. The outputs from this study will serve the automotive industry. This simulation tool will be beneficial in investment decision-making processes and research.

İlayda Demircioğlu
Özyeğin University · Institute of Graduate Studies in Science
2021
00
Master'sOpen AccessEN

Rüzgar santralleri için optimizasyon algoritmaları: Teori, tasarım ve uygulama

As the adaptation of wind power plants increases, questions regarding its optimal adaptation arise. This study answers two questions regarding wind farm optimization, where to place wind farms and how to arrange wind turbines inside the wind farm for maximum efficiency of the power plant. This study assesses the advantage of interdisciplinary studies involving computational intelligence and renewable energy fields. Computational intelligence often refers to using algorithms to solve optimization problems. Among the issues tackled in the optimization of wind plants are costs, sizing, and placement problems. The goal of this research is to increase the possibility of using renewable energy sources and reduce the dependence on fossil fuels globally. The environmental impact of wind farms on greenhouse gas reduction and employment opportunities are investigated. The findings suggest the attractiveness of the interdisciplinary approaches and the potential for computational intelligence and geographical information systems in transforming the wind energy sector. The study conducted also provides researchers a comprehensive review of available literature in addition to frameworks and models to help with subsequent studies on the optimization of wind farms.

Houda Zahmoul
Özyeğin University · Institute of Graduate Studies in Science
2021
00
Master'sOpen AccessEN

Elektrikli araç şarj cihazlarının dağıtım şebekesindeki yerleşiminin gerçek veri odaklı nicel araştırması

It is very evident that the number of electric vehicles worldwide will increase and that the dominant mobility concept in the future will be at the center of EVs. As a requirement of e-mobility, EVs should be rechargeable anywhere and in any situation, and this will increase EV-based electrification. The effects of EV charging stations on the grid are very important in the increasing electricity demand level and the installation of EV charging stations on distribution transformers (DTs) should be balanced and should not have negative effects on the transformers. This thesis aims to determine the quantitative deployments of EV charging stations to be installed on distribution transformers in an optimized and safe manner. The data of 8361 distribution transformers in the Central Anatolian Region were collected and analyzed. Data verification was completed with data cleansing methods and an effective EV charging station installation modeling was developed. Three different risk level modeling allows the installation of different EV charging stations in the most optimized way by determining the current state of the distribution transformer. The general situation of DTs in the region was determined with the pilot region application. Basic data can be revealed by modeling which region should be improved for the EV infrastructure to advance.

Algorithmic solutionsElectrificationElectric distribution systems+4
Hüseyin Tufan Güngör
Özyeğin University · Institute of Graduate Studies in Science
2022
00
Master'sOpen AccessEN

Olasılık ve istatistik temelli simülasyon ile elektrikli araç şarj istasyonlarının elektriksel ve finansal analizi

Increasing environmental pollution with the increase of vehicles that have internal combustion engines has led people to alternative methods due to the transportation sector's share in environmental pollution. Last decade, Researches have been increasıng sharply in this field to improve E-Mobility usage. On the other hand, automobile manufacturers who are producing internal combustion engines have announced that they will stop their R&D studies as of 2025. It is predicted that 300-350 million electric vehicles will be in traffic in the world by 2040. It is estimated that the energy demand created by this number of electric vehicles can be met with 350 million electric vehicle charging stations. In Turkey, it is estimated that 10 million electric vehicles will be on the roads in 2040. New charging stations will be established to meet the demand for electric vehicle charging stations that arise with the increase in electric vehicles. However, due to the lack of historical data, it is not entirely known how these stations will affect the electrification system in the region where they are installed in public areas. On the other hand, the number of stations does not increase rapidly due to the lack of a comparative example for investors who want to establish an electric vehicle charging station. In order to solve this problem, a simulation was created based on probabilistic and statistical analysis. As a result of the simulation, a realistic data set was created, and the effect of the new electric vehicle charging stations was discussed electrically and financially. There are two cases for using the simulation. One of the stations chosen for the simulation in the study is the shopping mall's car park, and another is a rest area on the Istanbul-Izmir highway.

Electric vehiclesElectrical analysisFinancial analysis+1
Doğukan Keser
Özyeğin University · Institute of Graduate Studies in Science
2022
00
Master'sOpen AccessEN

Farklı ihtiyaçlar için batarya boyutlandırma optimizasyon algoritmalarının modellenmesi

Battery energy storage systems (BESS) increase energy controllability and grid flexibility. One of the most important issues in BESS investments is optimal BESS sizing for various needs. In this thesis, it is aimed to develop an optimal battery sizing methodology for the consumer, producer and prosumer. By using Mixed-Integer Linear Programming and Mixed-Integer Quadratic Programming methods, optimal battery sizing algorithms that can be used by all end-user types for different purposes were developed. The advantage of this mathematical modeling is that it can be adapted for different scenario constraints with minor modifications. Various estimation algorithms were used to get more realistic results from the optimization algorithms for the future. Artificial neural network (ANN), deep neural network (DNN), and Long-Short Term Memory models were used to predict generation, consumption, and electricity market data. The importance of estimation algorithms in the smart grid ecosystem was emphasized and it was aimed to predict the needs for the future. Prediction methods and optimization algorithms were developed in the Python environment. Pandas, numpy, sklearn, keras, cvxpy libraries were actively used. It is hoped that it will be beneficial for the investments to be made within the scope of the smart grid concept.

Intelligence networksElectric power systemsEnergy storage system+2
Semanur Sancar
Özyeğin University · Institute of Graduate Studies in Science
2022
00
Master'sOpen AccessEN

Buzdolabı kablo tasarımının bozulma gücü testine deneysel etki analizi

The ever-increasing variety of electronic products brings with it many tests and controls. Especially for a product such as a refrigerator, which has an important place in the lives of users and works non-stop for almost its entire life, all compatibility tests that may affect the end-user must be carried out. Electromagnetic compatibility (EMC) tests are the most critical of these tests. In this study, the effect of the disturbance power test, one of the EMC tests, on the changing properties of the product was investigated. The effect of three different cable structures used in the refrigerator on the disturbance power test was investigated. These tests are carried out to cover 10 different frequencies in three products using different cable structures. Tables containing the oscilloscope results of these tests and the margin values corresponding to the frequencies are included in the document. In addition, visuals containing the differences between each cable structure and cable have been added. The results and analyzes of these tests are included in this thesis.

RefrigeratorElectromagnetic compatibilityElectronic equipments+1
Fırat Araz
Özyeğin University · Institute of Graduate Studies in Science
2022
10
Master'sOpen AccessEN

Elektrikli Bisiklet Bataryası için güvenlik test prosedürleri, standartları ve sonuçları

Recently, battery project studies have been increasing worldwide. Companies carrying out these projects must have their products checked for design safety and must obtain test approval reports in order to release their products on the market. IEC 62133-2 test approvals are required for lithium-ion rechargeable batteries, and EN 50604 test approvals are necessary for rechargeable li-ion batteries used in Light EV (electric vehicles). In addition, UN38.3(IEC 62281) battery transport tests must be applied for the transport of these batteries. These tests should be performed in an accredited laboratory, the safety of the design should be verified in this way, and then an accredited report should be issued to the product. There was no accredited laboratory in Turkey that performed tests in accordance with these standards. For this reason, companies were receiving services from overseas laboratories for the cell and battery approvals that they designed. There could be time and financial problems in overseas laboratories approvals. Due to these problems, an accredited laboratory was established at Vestel Electronics in Turkey, where tests can be carried out in accordance with IEC 62133-2, EN50604, and UN38.3 ( IEC 62281 ) standards. Since batteries contain BMS cards, their tests are also carried out in this laboratory. Within the scope of this accredited laboratory established in Turkey, it will be stated how the test setups that meet the test requirements are set up. Test results with these devices will be evaluated. The test results will belong to an E-bike battery project.

BatteryBicycleElectric vehicles+1
Mert Savrum
Özyeğin University · Institute of Graduate Studies in Science
2022
00
Master'sOpen AccessEN

Şeritte hareket eden bir robot için radyatif olmayan bir kablosuz güç aktarımı yöntemi teorisi ve deneyi

Today, the usage of robots is growing in the warehouses for automated organizing of goods and in the field for harvesting foods and other agriculture tasks. Many big resellers use robots in their distribution warehouses in an automated delivery system and, many manufacturers use them in their factories to transfer goods and components between production lines and warehouses integrated with enterprise resource planning (ERP) systems. A battery is needed such devices and, the battery size and capacity needed in such devices can be reduced thanks to WPT technology. It is useful for reducing environmental pollution resulted from disposing of batteries used for battery-operated vehicles and robots. With motivation for reducing battery size in robots, the main aim of the research is to find an alternative method of dynamic wireless power transfer to a moving object such as an electric vehicle or a robot that it moves on wheels on the floor. The well-known systems and their common series-connected/switched road embedded segment structures are discussed and challenged. Then, an alternative method of synchronization to the secondary resonance is presented. The suggested method eliminates switches between road-embedded coil units and makes the energy transition easy among units. This method shortens the length of the EMF radiative part of road embedded segments and the application is suitable for two-dimensional floor embedded structures for robot applications. The experimental and the simulation results are shared. The last chapter concludes thesis with discussion on the results of the thesis and future work.

Embedded structuresPower transferRobots
Mete Çeşmeci
Özyeğin University · Institute of Graduate Studies in Science
2022
00
Master'sOpen AccessEN

Current balancing circuit design for parallel LED strings in dual-transformer LLC resonant topology

New display technologies require small form factor designs with more backlight power. When multiple transformers drive parallel-connected LED strings, the difference in transformer leakage inductance and LED string voltage creates a backlight current imbalance. This current imbalance causes the following problems, - The low-voltage string draws more current. - High current causes hot spots on the backlight and a shorter lifespan. This issue is seen as an unevenly distributed backlight on the customer side. In this study, a cost-efficient balancing circuit for eight parallel-connected LED strings is simulated and developed utilizing two transformers in the half-bridge LLC resonant topology while providing the power in a comparatively slim design. The current imbalance for the 145W backlight power was 57.58% when there was an 8% change in leakage inductance between the transformers and a 10% change in LED string voltages. When the back cover was removed, the difference in temperature between the LED strings was recorded at room temperature as 31.8℃. The current imbalance between the LED strings is decreased to 9.99% when the current balancing components are connected to the circuit. When there is no difference between the LED string voltages, the maximum current imbalance is 2.59%. The cost of the proposed method is almost three times less than the conventional method, and for the same backlight power, it is 5.6% more efficient.

Switched mode power supplyEnergy efficiencyPower converters+1
Melisa Ersoy
Özyeğin University · Institute of Graduate Studies in Science
2022
00
Master'sOpen AccessEN

Piezoelektrik aktüatör tabanlı tarayıcı teknik tasarımı ve simülasyonu

Microelectromechanical systems (MEMS), offers compact and tailored solutions for a wide variety of applications such as laser beam guidance, tomography, microfluidic control, laser-based displays, LIDAR. These advantages of MEMS technology are important for the efficiency of the designed systems. This thesis aims to develop a scanner with high scanning frequency and scanning angle moved by a PZT-4 thin film. Our thesis focuses on optimizing the designed system using finite element model- ing. With this modeling, the movements resulting from the electrical input applied to the PZT-based actuator are evaluated. Additionally, with finite element modeling, the stress distribution on the system and its natural mode oscillations are examined. In out-of-phase motion, the system was operated with 100V at a mechanical res- onance frequency of 91 kHz. In-phase movement is at 208 Hz levels. This movement adds depth to the scan. Under these operating conditions, a scanning angle of 12.5 degrees is achieved. The width of the mirror is made is 0.66 mm. Mechanical defor- mations occur as a result of the system operating at high frequency. The theoretical calculation shows that 70 % of the existing mirror surface can be used efficiently. According to theoretical calculations, the power consumption of the system is around 1.5 mW. The designed system is intended to be used in applications where speed is needed, such as laser projectors and medical applications.

Uğur Çelik
Özyeğin University · Institute of Graduate Studies in Science
2024
00
Master'sOpen AccessEN

Pandemi sonrası Türkiye gün öncesi piyasasında doğal gaz fiyatlarının elektrik fiyatlandırması üzerindeki gecikmeli etkilerinin analizi

This study investigates the correlation between natural gas and electricity prices in Türkiye, which has experienced significant spikes in natural gas prices due to the pandemic and regional conflicts. The determination of reference prices for electricity and natural gas in Türkiye is facilitated by transparent markets operated by the Energy Exchange. Using data from the EXIST transparency platform, this research employs regression techniques to determine the temporal relationship between abrupt increases in natural gas prices and their impact on electricity prices. Prior studies have shown that elevated natural gas prices affect electricity prices with a delay, depending on the proportion of natural gas in a country's energy supply. This research applies a similar methodology to assess the time-lag effect of natural gas prices on electricity prices within the Turkish energy market post-COVID-19. Metrics such as mean absolute error, mean square error, mean absolute percent error, and mean percent error validate the findings. The results demonstrate a eightyfive-day time lag in the correlation between natural gas and electricity prices in Türkiye. This delay improves forecasting accuracy by up to 16.17% in terms of error reduction.

Betül Elif Özmen
Özyeğin University · Institute of Graduate Studies in Science
2024
00
Master'sOpen AccessEN

Paralel rezonans konvertör ile akim piklerini önlemek için düşük gürültü blok güç devresi tasarimi

Low Noise Block (LNB) circuits are an essential component of satellite receiving systems in televisions. LNB converts high-frequency satellite signals into a lower frequency range for transmission to the receiver. Also, LNB ensures low-noise signal reception, enables polarization control, supports multiple frequency bands, and amplifies signals. The LNB power circuits provide power to the LNB. Voltage-controlled boost converters are commonly used in these circuits; the current limiting feature used in these circuits can be insufficient to handle sudden current spikes, leading to feedback from end users. They are not immune to short circuits because of installation problems with satellite cables, which are a common failure mode that potentially damages the LNB power or other components in the television, resulting in a loss of satellite signal. Hence, leading to a poor user experience and harming the manufacturer's reputation. Replacing a damaged LNB power circuit is also difficult because it is often integrated into the TV's design and is not easily replaceable. To mitigate this problem, a parallel resonant converter approach is introduced that limits the output current up to a pre-defined level by frequency control. Parallel resonant converter circuits operate as current sources so maximum current cannot exceed this defined operation frequency point under any circumstances. In this study, the mathematical equations of the proposed circuit were obtained, and then the operating conditions were determined. The operation frequency ranges were interpreted graphically using Matlab, and the parallel resonant converter design was evaluated using a simulation model by using MATLAB's Simulink Toolbox. These results show that the parallel resonant converter approach maintains a stable current and reduces the risk of short-circuit damage. The proposed design is expected to reduce repair costs due to LNB power circuits for TV manufacturers.

Cansel Şahin Ceter
Özyeğin University · Institute of Graduate Studies in Science
2025
00

Other supervisors