Geri Dön

İnsansız hava araçları için batarya-yakıt hücresi hibrit enerji yönetim sisteminin geliştirilmesi

Development of a battery-fuel cell hybrid energy management system for unmanned aerial vehicles

PDF İndir

  1. Tez No: 992443
  2. Yazar: BARAN KUL
  3. Danışmanlar: DOÇ. DR. MEHMET ONUR GÜLBAHÇE
  4. Tez Türü: Yüksek Lisans
  5. Konular: Elektrik ve Elektronik Mühendisliği, Electrical and Electronics Engineering
  6. Anahtar Kelimeler: Belirtilmemiş.
  7. Yıl: 2026
  8. Dil: Türkçe
  9. Üniversite: İstanbul Teknik Üniversitesi
  10. Enstitü: Lisansüstü Eğitim Enstitüsü
  11. Ana Bilim Dalı: Elektrik Mühendisliği Ana Bilim Dalı
  12. Bilim Dalı: Elektrik Mühendisliği Bilim Dalı
  13. Sayfa Sayısı: Belirtilmemiş.

Özet

İnsansız hava araçlarının özellikle multirotor konfigürasyonlarda kullanımının yaygınlaşması, uçuş süresi, görev menzili ve enerji verimliliği gibi başarım kriterlerini kritik hale getirmiştir. Günümüzde yaygın olarak kullanılan lityum batarya tabanlı enerji sistemleri, yüksek güç yoğunluğu ve hızlı dinamik yanıt avantajlarına sahip olmakla birlikte, sınırlı özgül enerji kapasiteleri nedeniyle uzun süreli uçuş görevlerinde yetersiz kalmaktadır. Bu bağlamda hidrojen yakıt hücreleri, yüksek özgül enerji ve uzun çalışma süreleri sunmaları nedeniyle İHA uygulamaları için önemli bir alternatif olarak öne çıkmaktadır. Ancak yakıt hücrelerinin yavaş dinamikleri, ani yük değişimlerine karşı sınırlı tepkileri ve sistem yaşlanmasına duyarlılıkları, tek başına kullanımlarını zorlaştırmaktadır. Bu nedenle, yakıt hücrelerinin bataryalar ile birlikte hibrit bir yapı içerisinde kullanılması, her iki enerji kaynağının avantajlarını bir araya getiren etkili bir çözüm olarak değerlendirilmektedir. Bu tez çalışmasında, multirotor İHA'lar için hidrojen yakıt hücresi ve lityum bataryadan oluşan aktif hibrit bir güç sistemi tasarlanmış, MATLAB/Simulink ortamında modellenmiş ve ayrıntılı olarak benzetimi yapılmıştır. Çalışmanın temel amacı; batarya destekli bir hibrit yapı ile yakıt hücresinin dinamik sınırlamalarını telafi eden, uçuş süresini uzatan ve enerji verimliliğini artıran bir güç sistemi ve buna ait enerji yönetim stratejisi geliştirmektir. Bu kapsamda yakıt hücresi, batarya, güç elektroniği dönüştürücüleri, enerji yönetim sistemi ve kontrol mekanizmaları bütünleşik bir yapı içerisinde ele alınmıştır. Tez kapsamında öncelikle proton değişim membranlı (PEM) yakıt hücresinin matematiksel ve fiziksel modelleri ayrıntılı biçimde incelenmiştir. Yakıt hücresinin gerilim karakteristiği; tersinir gerilim, aktivasyon, ohmik ve konsantrasyon kayıpları dikkate alınarak modellenmiş; gaz dinamiği, kısmi basınçlar ve Faraday yasasına dayalı fiziksel denklemler kullanılarak sistemin elektrokimyasal davranışı temsil edilmiştir. Bu modelleme, yakıt hücresinin verimli çalışma bölgelerinin belirlenmesine ve enerji yönetim stratejisinin oluşturulmasına temel teşkil etmiştir. Güç elektroniği katmanında, yakıt hücresinin 24 V seviyesindeki çıkış gerilimini 48 V doğru akım (DA) bara gerilimine yükseltmek amacıyla 5-fazlı faz kaydırmalı yükseltici (interleaved boost) dönüştürücü tasarlanmıştır. Çok fazlı yapı sayesinde giriş ve çıkış akım/gerilim dalgalanmaları azaltılmış, yüksek güç seviyelerinde daha verimli ve dengeli bir çalışma elde edilmiştir. Batarya tarafında ise hem şarj hem de deşarj işlemlerinin aynı donanım üzerinden gerçekleştirilebilmesi için 4-anahtarlı çift yönlü alçaltıcı-yükseltici (buck–boost) dönüştürücü kullanılmıştır. Her iki dönüştürücü için detaylı tasarım denklemleri verilmiş, indüktör ve kapasitör değerleri belirlenmiş ve kapalı çevrim gerilim–akım kontrol yapıları kompanzatörler kullanılarak modellenmiştir. Enerji yönetim sistemi tasarımı kapsamında, hibrit enerji sistemlerinde kullanılan pasif ve aktif mimariler incelenmiş, bu çalışma için aktif mimarinin daha uygun olduğu gösterilmiştir. Ardından İHA uygulamalarında kullanılan farklı EMS yaklaşımları; kural tabanlı kontrol, frekans tabanlı yöntemler, eşdeğer yakıt tüketimi minimizasyonu (ECMS), model öngörülü kontrol (MPC) ve dinamik programlama açısından değerlendirilmiştir. Gerçek zamanlı uygulanabilirlik, hesaplama maliyeti, sistem güvenilirliği ve kontrol edilebilirlik kriterleri dikkate alındığında, bu çalışma kapsamında kural tabanlı bir enerji yönetim sisteminin tercih edilmesi uygun bulunmuştur. Kural tabanlı EMS, yakıt hücresinin verimli çalışma ve mümkün olduğunca sabit güç seviyelerinde çalışmasını sağlayacak şekilde tasarlanmıştır. Hidrojenin batarya enerjisine kıyasla daha maliyetli olması, yakıt hücresinin yaşlanma ve yıpranma etkileri ile bataryanın şarj durumu (SoC) kısıtları, kural setlerinin oluşturulmasında temel kriterler olarak ele alınmıştır. Talep edilen güç seviyesi ve batarya şarj durumu (SoC) bilgisine bağlı olarak yakıt hücresi çıkış gücü belirlenmiş, kalan güç ihtiyacı batarya tarafından karşılanmıştır. Ayrıca, sistemin ilk çalıştırma anı ve ani geçici rejimler için frekans tabanlı bir anahtarlama mekanizması kullanılarak bataryanın öncelikli olarak devreye alınması sağlanmıştır. Kural tabanlı yapı, daha esnek ve yumuşak bir kontrol sağlamak amacıyla bulanık mantık kontrolcü ile gerçekleştirilmiştir. Talep gücü ve batarya şarj durumu (SoC) girişleri kullanılarak yakıt hücresi çıkış gücü bulanık mantık ile belirlenmiştir. Bulanık mantık kontrolcüsünün üyelik fonksiyonları ve eşik değerleri, hidrojen tüketimi ve görev sonu hedeflenen SoC değerinden sapmasını içeren çok kriterli bir maliyet fonksiyonu kullanılarak genetik algoritma (GA) ile optimize edilmiştir. Optimizasyon süreci çevrim dışı olarak yürütülmüş ve elde edilen en iyi parametre seti ile kontrolcü nihai haline getirilmiştir. Geliştirilen bütünleşik sistem modeli, uçuş sırasında talep edilecek zamana bağlı yük profilleri altında benzetimi yapılmıştır. Elde edilen sonuçlar, yüksek SoC ve yüksek güç taleplerinde batarya ağırlıklı bir güç paylaşımı gerçekleştirildiğini; SoC değerinin nominal aralığa yaklaşmasıyla birlikte yakıt hücresinin verimli ve sabit güç bölgelerinde devreye alındığını göstermiştir. Ani ve dalgalı yük taleplerinde bataryanın hızlı dinamik yanıtından faydalanılmış, böylece yakıt hücresinin gereksiz güç dalgalanmalarına maruz kalması önlenmiştir. Sonuç olarak, hidrojen tüketimi açısından maliyet etkin, yakıt hücresi ömrünü koruyan ve uçuş süresini artıran bir hibrit güç sistemi elde edilmiştir.

Özet (Çeviri)

The rapid proliferation of Unmanned Aerial Vehicles (UAVs) in diverse sectors, ranging from surveillance and mapping to logistical transport, has underscored the critical need for advanced energy systems that offer both high energy density and robust dynamic response. While multirotor UAVs provide superior maneuverability and operational flexibility, their performance is fundamentally constrained by the energy capacity and reliability of their power sources. Current battery-based systems, though capable of meeting high power demands during transients, often fail to support long-endurance missions due to their inherent energy density limitations. In this context, Proton Exchange Membrane Fuel Cells (PEMFCs) have emerged as a promising alternative, offering high specific energy density and environmental benefits. However, the primary challenge hindering their standalone application is the slow dynamic response of PEMFCs during rapid load fluctuations—such as those encountered during takeoff or aggressive maneuvers—which can lead to critical fluctuations in the DC bus voltage and threaten mission safety. To address these limitations, this thesis proposes an active hybrid energy system that integrates the high energy density of PEMFCs with the high power density and rapid dynamic response of lithium batteries. The primary objective of the study is to develop a holistic engineering framework comprising a detailed mathematical model of the PEMFC, the design and modeling of appropriate DC-DC converters, and the implementation of an advanced Energy Management System (EMS). This system is intended to maintain stability and reliability under various dynamic flight conditions. A central component of this research is the development of a comprehensive mathematical model of the PEMFC, capturing its electrochemical, thermodynamic, and physical behaviors. This non-linear model is based on the principle of subtracting activation, ohmic, and concentration losses from the ideal thermodynamic Nernst potential. The modeling process incorporates essential physical parameters, including the partial pressures of hydrogen and oxygen, temperature variations, and membrane hydration. Specifically, the Nernst equation is utilized to calculate the reversible open-circuit voltage as a function of operating conditions. Furthermore, electrochemical kinetics, such as activation losses and Faraday's Law for gas consumption, are modeled to ensure a realistic representation of the fuel cell's dynamic response. The model also accounts for critical factors like water vapor saturation pressure, which directly influences reactant partial pressures and overall cell efficiency. In order to ensure compatibility at the system level and enable the effective implementation of energy management strategies, both sources were designed to operate at a nominal voltage level of 24 V. However, the common DC distribution bus supplying the propulsion system and onboard electronics of the unmanned aerial vehicle (UAV) was specified at 48 V. This voltage difference necessitated the use of appropriately designed power electronic converters to interface each energy source with the DC bus. Accordingly, two different DC–DC converter topologies were developed. These converters not only provide voltage level matching but also enable controlled power flow and improved system stability. PEM fuel cells are characterized by low output voltage, high current levels, and limited dynamic response capability. These characteristics make direct connection to a high-voltage DC bus unsuitable. Therefore, a boost-type DC–DC converter operating in continuous conduction mode (CCM) was selected to elevate the fuel cell voltage to the required DC bus level while minimizing current ripple. To overcome the limitations of single-phase boost converters at high power levels, a five-phase interleaved boost topology was adopted. In this structure, multiple identical boost phases operate in parallel with evenly phase-shifted switching signals. This interleaving approach distributes the input current among the phases, significantly reducing current stress on individual semiconductor devices. In addition, phase-shifted operation results in partial cancellation of ripple components, leading to reduced input and output current ripple. The converter was designed for a nominal voltage conversion from 24 V to 48 V, corresponding to a duty cycle of approximately 0.5. A switching frequency of 20 kHz was selected as a compromise between switching losses and passive component size. Operation in CCM was intentionally maintained to ensure smooth input current, which is critical for fuel cell durability. Inductor values were determined based on strict current ripple limitations, targeting a ripple level of 1.5–2% of the nominal current. As a result, identical inductors with a value of 0.5 mH were selected for each phase. The DC bus voltage ripple was limited to below 1%, and an output capacitance of 2 mF was chosen to ensure sufficient energy buffering under dynamic load conditions. To achieve stable and robust operation, a dual-loop closed-loop control architecture was implemented. The outer voltage loop regulates the DC bus voltage by comparing the measured voltage with a 48 V reference. The resulting error is processed by a proportional–integral (PI) controller, which generates a reference value for the total input current. The inner control layer consists of individual current PI controllers for each phase. The total reference current is evenly distributed among the phases, and each phase current is regulated independently. This structure ensures accurate current sharing and prevents phase imbalance. Phase-shifted PWM signals with 72° separation are generated to maintain interleaved operation. The control gains were selected through simulation-based tuning, with the voltage loop bandwidth intentionally kept lower than that of the current loops to preserve hierarchical control behavior. The lithium-ion battery serves as both an auxiliary power source and an energy storage element in the hybrid system. Therefore, bidirectional power flow between the battery and the DC bus is required. To meet this requirement, a four-switch bidirectional buck–boost converter topology was selected. During battery discharge, the converter operates in boost mode, increasing the battery voltage from 24 V to the 48 V DC bus. Conversely, during battery charging, the converter operates in buck mode, reducing the DC bus voltage to a level suitable for battery charging. The four-switch topology enables seamless transitions between operating modes through appropriate switching control, eliminating the need for additional mechanical switches or diode-based power paths and thereby improving efficiency. The converter was designed to operate at a switching frequency of 100 kHz in order to reduce passive component size. A single shared inductor was employed for both operating modes, and its value was determined to ensure CCM operation under worst-case ripple conditions. Based on analytical calculations, an inductance of 0.4 mH was selected. Capacitor values were chosen to maintain voltage ripple below 1% on both the DC bus and battery sides. Different control strategies were implemented for the two operating modes of the bidirectional converter. In discharge (boost) mode, the primary objective is to regulate the DC bus voltage under varying load conditions. A PI controller was therefore employed to eliminate steady-state error and ensure stable voltage regulation. In charging (buck) mode, precise control of the battery charging voltage and suppression of transient current spikes are critical to battery health. For this reason, a proportional–integral–derivative (PID) controller was adopted to improve transient response and damping characteristics. The controller parameters were determined through iterative simulation to ensure smooth and stable charging behavior. Hybrid energy systems are introduced as multi-source configurations designed to improve power continuity, efficiency, and system reliability, particularly under fluctuating load conditions. Fuel cell–battery hybrid systems are emphasized due to their complementary characteristics, wherein fuel cells provide high energy density and sustained power, while batteries offer high power density and rapid dynamic response. However, it is stressed that system performance is not solely dependent on the individual energy sources, but also on the adopted architectural topology and the applied energy management strategy. Accordingly, hybrid architectures are categorized into passive and active configurations. Passive hybrid architectures are described as systems in which energy sources are directly connected to a common DC bus without the use of active power electronic interfaces. In such systems, power sharing is inherently determined by the internal resistances, voltage levels, and natural electrical characteristics of the sources. Although passive architectures offer advantages such as structural simplicity, low cost, reduced weight, and the absence of converter losses, they suffer from a critical lack of controllability. In particular, fuel cells are exposed to sudden load transients, which may accelerate degradation and compromise system stability. Moreover, DC bus voltage regulation is weak and highly dependent on load conditions, rendering passive architectures unsuitable for UAV applications that require precise voltage control and rapid transient handling. Active hybrid architectures, by contrast, employ DC–DC power converters between energy sources and the DC bus, enabling decoupled voltage levels and actively controlled power flow. Among the various active configurations, fully active topologies—utilizing a unidirectional converter for the fuel cell and a bidirectional converter for the battery—are identified as providing the highest level of flexibility and controllability. In this structure, an EMS supervises power sharing by issuing reference commands to the converters based on real-time system states and load demands. The advantages of this architecture include stable DC bus voltage regulation, protection of the fuel cell against high-frequency load variations, extended battery lifetime through controlled charge–discharge cycles, and seamless integration of sources with different voltage levels. These benefits are achieved at the expense of increased system complexity, weight, cost, and converter-related power losses. Nevertheless, for UAV missions requiring high reliability and precise power control, the fully active hybrid architecture is justified and therefore selected in this study. The primary objectives of the EMS are defined as maintaining DC bus voltage stability, satisfying transient power demands, protecting the fuel cell from rapid dynamics, preserving battery SoC within safe limits, and maximizing overall system efficiency. Various EMS strategies reported in the literature are reviewed, including rule-based control, frequency-based methods, Equivalent Consumption Minimization Strategy (ECMS), Model Predictive Control (MPC), and Dynamic Programming (DP). Rule-based EMS is identified as a suitable approach for real-time UAV applications due to its low computational burden, transparency, and reliability. In this method, power sharing decisions are governed by predefined logical rules derived from engineering expertise and system constraints. While rule-based strategies do not guarantee global optimality, their robustness and real-time feasibility are considered more critical for safety-oriented UAV missions. Frequency-based EMS methods, which allocate low-frequency power components to the fuel cell and high-frequency components to the battery, are acknowledged for their natural protection of fuel cells but criticized for their sensitivity to filter design and inability to directly manage battery SoC. ECMS is discussed as a near-optimal approach that minimizes instantaneous equivalent fuel consumption; however, its dependency on equivalence factors and increased computational requirements limit its practicality for UAVs. Advanced optimization-based methods such as MPC and DP are recognized for their theoretical optimality but are deemed impractical due to high computational demands, model dependency, and lack of real-time feasibility. Based on these considerations, a rule-based EMS supported by transient detection mechanisms is adopted. The selected EMS architecture prioritizes stable fuel cell operation at predefined power levels while assigning transient and peak power demands to the battery. Special attention is given to system start-up and sudden load increase scenarios, during which the fuel cell is temporarily decoupled from the DC bus to prevent voltage collapse. This hybrid EMS structure enhances DC bus voltage stability and ensures deterministic system behavior, which is essential for UAV flight safety. The EMS decision-making framework is constructed using two primary inputs: demanded load power and battery SoC. Both variables are classified into three linguistic levels (low, medium, high), resulting in a comprehensive rule base that determines the fuel cell operating mode (minimum, nominal, or maximum power). This rule set ensures that the fuel cell operates within its optimal efficiency and durability range, while the battery functions as a buffer to absorb dynamic load variations. The resulting power balance is achieved by allocating the residual load demand to the battery. To mitigate abrupt transitions inherent in conventional rule-based control, a fuzzy logic controller (FLC) is introduced. The FLC employs Mamdani-type inference with trapezoidal and triangular membership functions to achieve smooth and continuous control behavior. The controller inputs consist of demanded power and battery SoC, while the output corresponds to the fuel cell reference power. Membership function parameters are initially defined based on engineering judgment and system constraints, enabling gradual transitions between operating modes and reducing stress on system components. Recognizing that manually tuned fuzzy parameters may not yield optimal system performance, an optimization framework is subsequently introduced. The EMS optimization problem is formulated as a nonlinear, multi-objective task aimed at minimizing hydrogen consumption while achieving a desired terminal battery SoC. Various optimization techniques are evaluated, including Particle Swarm Optimization, Simulated Annealing, deterministic methods, and Genetic Algorithms (GA). GA is ultimately selected due to its suitability for discontinuous, nonlinear parameter spaces and its ability to handle multiple performance objectives without requiring gradient information. The GA-based optimization is performed offline in a MATLAB environment, targeting key fuzzy membership function parameters and fuel cell power thresholds. A weighted cost function combining total hydrogen consumption and deviation from target end-of-mission SoC is minimized. Optimization results demonstrate that the GA-refined EMS parameters significantly improve energy utilization efficiency while maintaining safe battery operation. The robustness and repeatability of the optimization outcomes further validate the effectiveness of the proposed approach. In conclusion, this chapter presents a systematically developed hybrid fuel cell–battery power architecture and an optimized fuzzy logic–based EMS for UAV applications. By integrating rule-based decision logic, fuzzy control, and genetic algorithm optimization, the proposed system achieves a balanced trade-off between performance, reliability, and real-time feasibility. The final EMS design provides a scalable and extensible foundation for future enhancements, including adaptive and learning-based energy management strategies.

Benzer Tezler

  1. Tez No

    723995

    Bir İHA'nın üst sezgisel algoritmalarla enerji yönetimi ve eniyilemesi

    Energy management and optimization based on meta-heuristic algorithms for a UAV

    HASAN ÇINAR

    Doktora

    Türkçe

    Türkçe

    2022

    Uçak MühendisliğiGebze Teknik Üniversitesi

    Makine Mühendisliği Ana Bilim Dalı

    PROF. DR. İLYAS KANDEMİR

  2. Tez No

    1000551

    Hibrit güç aktarımlı insansız hava aracının matlab/simulink modeli ve enerji yönetimi

    Matlab/simulink modeling and energy management of a hybrid powered unmanned aerial vehicle

    AYSEL GÖKÇEBUZHAN

    Yüksek Lisans

    Türkçe

    Türkçe

    2026

    Havacılık ve Uzay MühendisliğiFırat Üniversitesi

    Havacılık Bilimi ve Teknolojileri Ana Bilim Dalı

    DR. ÖĞR. ÜYESİ BUĞRA YILMAZ

  3. Tez No

    813324

    Hibrit insansız hava araçlarına (iha) yönelik jeneratör sistemleri için batarya yönetim sistemi tasarımı

    Battery management system design for the generator systems intended to hybrid unmanned aereal vehicles

    MEHMET YOZGAT

    Yüksek Lisans

    Türkçe

    Türkçe

    2023

    Elektrik ve Elektronik MühendisliğiYozgat Bozok Üniversitesi

    Elektrik-Elektronik Mühendisliği Ana Bilim Dalı

    DR. ÖĞR. ÜYESİ EMRAH ÇETİN

  4. Tez No

    888646

    Modelling and simulation of the pem fuel cell on mq1- predator unmanned air vehicle

    Mq-1 predator insansız hava aracında pem tipi yakıt pilinin modelleme ve simülasyonu

    OSMAN SÖZEN

    Yüksek Lisans

    İngilizce

    İngilizce

    2024

    Elektrik ve Elektronik Mühendisliğiİstanbul Teknik Üniversitesi

    Enerji Bilimi ve Teknolojileri Ana Bilim Dalı

    PROF. DR. İLHAN KOCAARSLAN

  5. Tez No

    10819

    Los Personajes y los temas en las novelas de unamuno

    Başlık çevirisi yok

    HALE GÖKNAR

    Doktora

    İspanyolca

    İspanyolca

    1989

    DilbilimAnkara Üniversitesi

    İspanyol Dili ve Edebiyatı Ana Bilim Dalı

    PROF. DR. MUZAFFER ARIKAN

Geri Dön