서지주요정보
Combustion dynamics of lean-premixed axial fuel-staged combustor: self-excited instability, flame transfer function, and low-order modeling = 희박 예혼합 다단 연소기의 연소동역학: 자발 불안정, 화염전달함수 및 저차원 모델링
서명 / 저자 Combustion dynamics of lean-premixed axial fuel-staged combustor: self-excited instability, flame transfer function, and low-order modeling = 희박 예혼합 다단 연소기의 연소동역학: 자발 불안정, 화염전달함수 및 저차원 모델링 / Yongseok Choi.
발행사항 [대전 : 한국과학기술원, 2025].
Online Access 원문보기 원문인쇄

소장정보

등록번호

8043573

소장위치/청구기호

학술문화관(도서관)2층 학위논문

DAE 25013

휴대폰 전송

도서상태

이용가능(대출불가)

사유안내

반납예정일

리뷰정보

초록정보

Advanced gas turbine engines are expected to play a crucial role as a stabilizer in the rapid, widespread electrification and renewable transformation of energy systems. Axial staged combustion methods, whether for newly developed or retrofitted gas turbine combustors, are promising candidates for enhancing combined cycle efficiency by more than 65$\%$ without significantly increasing nitric oxide emissions. In this concept, the majority of the fuel is supplied to the primary reaction region operating under lean-premixed conditions, and the remaining fuel/air mixtures are then injected into the vitiated crossflow near the combustor exit to increase the turbine inlet temperature. However, combustion instabilities under axial-fuel-staging conditions can become more complex than those in conventional single-flame combustion, due to the presence of multiple discrete reaction zones and their various interactions. A proper understanding of these dynamics is essential for the successful application of the technique. Driven by both physical and practical motivations, this doctoral dissertation intensively addresses the combustion dynamics of a lean-premixed axial fuel-staged system, in an attempt to significantly expand our understanding of these phenomena. Using a lean-premixed axially staged combustor, several experimental investigations are conducted under various operating conditions, which are meticulously designed for systematic examination of the dynamics. During the investigations, thermophysical data regarding self-excited conditions, related flame dynamics, and forced responses are measured using various methods; these include dynamic pressure measurements, photomultiplier tubes, hot-wire anemometers, high-speed OH* chemiluminescence and OH planar laser-induced fluorescence image measurements, and acoustic external forcing. Numerical and analytical approaches---specifically finite element methods-based and reduced-order model-based linear acoustic calculations---are also carried out to support empirical findings. Initially, the study investigates the link between first- and second-stage flame dynamics in the lean-premixed axially staged combustor with various inlet conditions. The findings indicate that the second-stage jet-in-vitiated-crossflow flame is preferentially coupled to higher-order acoustic modes than the upstream primary flame with the same characteristic nozzle dimension. The excitations driven by each flame appear to be mutually incompatible, with this selective behavior critically dependent on the secondary flame's position relative to the acoustic mode shape. When the second-stage flame is located near a pressure node or a velocity antinode, the dynamics of the flame feature conspicuous flapping motion in the crossflow direction due to the influence of the primary flame-induced crossflow velocity fluctuations. On the other hand, when located near a pressure antinode, the secondary jet flame periodically exhibits combined dynamics of jet merging-induced flame front annihilation and the growth of non-axisymmetric coherent structures. In this case, high-intensity pressure oscillations are sustained by the local heat release fluctuations generated solely from the transverse reacting jets, while the upstream primary flame remains nearly unperturbed, or decoupled from the underlying feedback processes. In addition, to understand the synchronized or isolated dynamics of the system, the $\rm H_{2}$ concentration in the primary flame was varied without altering the secondary operating conditions. The results reveal two new phenomena: the simultaneous excitation of both flames sharing the same acoustic mode under higher $\rm H_{2}$ conditions, and instabilities induced by the secondary flame in isolation, without noticeable modulations in the primary flame. The following study investigates the distinctive characteristics of the transfer function of transverse reacting jets in vitiated crossflow, fundamentally important yet largely unknown. During the research, reduced-order modeling combined with two different flame transfer functions (for the primary and secondary flames)—an approach that has not been previously demonstrated—is performed to predict the stability of the axially staged system. From detailed measurements of transfer functions of lean-premixed primary and secondary flames in response to harmonic velocity disturbances, the results show that the transfer function of the second-stage flame is described as having relatively larger gain without undulating patterns, near-linear slow decay in magnitude with respect to the forcing frequency, and remarkably shorter duration response time. Using the reduced-order modeling approach that considers the crucial finite time delay between two distinct transfer functions, validated against empirical data, we demonstrate that self-induced instabilities in the axially staged system can be driven simultaneously by the dynamics of both primary and secondary flames. Additionally, the triggering of second-stage-induced instability is characteristically connected to higher-order acoustic modes, consistent with the aforementioned results. Such non-axisymmetric intra-combustor interactions generate a cascade of spectral peaks, including combinations of the first and third longitudinal modes and their intermodulation frequencies. Here, the secondary jet flame exhibits more complicated modal dynamics, manifested as the L3 mode-coupled jet merging-related flame surface annihilation and the L1 mode-coupled lateral movements of the transverse jet column. Additionally, the analysis reveals a close relationship between the amplitude of the secondary flame's fluctuations and the direction of the crossflow. Results from the consecutive experimental investigations are remarkable, revealing previously unknown findings about the combustion dynamics of the axially staged combustor and elucidating the underlying mechanisms of the complex thermoacoustic interactions in the two lean-premixed flames. The current integrated empirical and analytical findings are projected to serve as both foundational knowledge and an effective tool for the future development and operation of axial fuel-staged gas turbine engines.

가스터빈은 현재 에너지 산업에서의 빠르고 광범위한 신재생 에너지로의 전환(renewable transformation)과 그에 따른 전기화(electrification) 과정에서 공급망 안정화에 중요한 역할로 주목받고 있다. 최근 개발 및 개선되고 있는 가스터빈 연소기에 적용되는 다단 연소 기법은 질소 산화물의 추가 생성을 억제하면서 65$\%$ 이상의 복합 효율 향상을 이룰 수 있는 유력한 방법이다. 해당 기법이 적용된 연소기에서는 희박 조건으로 주 연소 영역을 운용하고, 나머지 연료/공기 혼합물을 연소기 출구 근처에서 높은 온도의 연소 후 생성물에 분사하여 터빈 입구 온도를 올리는 방식이다. 하지만 이러한 다단 연소 조건에서는 연소실 내에 여러 열원이 존재하며, 이들 간의 상호작용으로 인해 단일 연소 조건보다 더 복잡한 형태의 연소불안정이 발생할 수 있다. 이 연소불안정에 대한 이해는 다단 연소 기술 적용과정에서 필수 전제조건이다. 본 학위논문은 물리적 및 실용적 동기에서 출발하여, 희박 예혼합으로 운용되는 축방향 연소 시스템의 연소불안정 현상을 중점적으로 다루며 이러한 현상에 대한 이해를 넓히는 것을 목표로 하였다. 본 연구에서는 다단 연소기 내에서 발생하는 연소동역학을 체계적으로 이해하기 위해 실험 조건을 세심하게 구상하였고, 희박 예혼합 축방향 다단 연소기를 이용하여 진행되었다. 연구 과정에서 압력 섭동 계측, 광증배관, 열선 유속계, 고속 OH* 자발광 및 OH 평면 레이저 유도 형광 이미지 계측, 외부 음향 가진 등의 다양한 계측 및 실험 기법을 이용하여 자발 불안정과 해당 화염 동역학, 그리고 강제 응답 조건에서의 열물리적 특성을 측정하였다. 또한, 유한 요소법 (finite element method) 과 축소차수모델(reduced order model) 기반의 선형 음향 분석을 수행하여 실험 결과를 이해하는데 도움이 될 수 있도록 하였다. 먼저, 다단 연소 시 1차와 2차 연소실의 화염 동역학 사이 발생하는 상호작용을 확인하기 위해, 희박 예혼합 다단 연소기에서 다양한 연료 공급 조건을 설정해 실험을 수행하였다. 실험 결과, 2차 연소실의 횡단류 제트 유동 형태의 화염은 동일한 노즐 규격임에도 불구하고, 상류의 1차 연소실 화염보다 더 높은 음향 모드와 우선적으로 결합하여 불안정을 일으키는 것을 확인하였다. 각 화염에 의해서 발생한 불안정은 상호 양립하지 않은 형태로 나타났으며, 이렇게 선택적으로 발생하는 불안정은 연소기 음향장 내에서 2차 화염의 상대적인 위치에 따라 영향을 받아 크게 달라지는 것으로 관찰되었다. 2차 화염이 음향장의 압력 마디(node), 혹은 속도 섭동의 배(antinode) 부근에 위치하는 경우, 2차 화염은 1차 화염 불안정에 의해 발생하는 유동 섭동에 의해서 횡단류 방향, 즉 좌우로 휘날리는 거동을 보였다. 반면, 2차 화염이 압력 섭동의 배 근처에 위치하는 경우, 2차 횡단류 수직 분사 화염은 제트 병합 거동으로 인한 화염면의 급격한 소염과 비대칭 구조의 거대 제트 화염 형성이 혼재된 거동를 겪으며, 이러한 움직임을 주기적으로 나타냈다. 이 과정에서 발생하는 2차 제트 화염에서의 국부적인 열 방출 섭동이 강한 압력 섭동을 발생 및 유지할 수 있는 것으로 관측되었다. 이때 상류의 1차 화염은 열음향학적으로 거의 안정 상태로 미약한 섭동을 보이는 것으로 확인되었다. 추가로 두 화염의 동조 혹은 독립된 불안정 거동을 이해해보고자 2차의 운용 조건은 유지한 상태로, 1차 화염의 수소 몰분율 변화에 따른 연소불안정 거동을 확인하였다. 이 결과 두 가지 새로운 현상이 관찰되었는데, 하나는 높은 수소 분율 조건에서 1차 화염과 2차 화염이 동일한 주파수로 동시에 불안정을 일으키는 것이고, 다른 하나는 1차 화염이 거의 섭동하지 않아도 2차 화염이 독립적으로 불안정을 일으킬 수 있다는 것이다. 이어지는 연구에서는 아직 많은 분석이 이루어지지 않은 횡단류 제트 유동 형태의 화염에서, 화염 동역학을 이해하기 위한 중요한 방법으로 화염 전달 함수의 특성을 파악하고자 하였다. 이 연구에서는 다단 연소의 1차와 2차 화염에서 계측된 서로 다른 화염전달함수를 축소차수모델(reduced order model)에 동시에 적용하여 계산하는 새로운 접근 방식을 고안하였고, 이를 연소 시스템의 안정성 파악에 활용하였다. 외부 유동 가진을 통한 각 화염의 전달함수 계측을 통해 2차 횡단류 수직 분사 화염의 화염전달함수는 상대적으로 높은 증폭을 보이며, 주파수 증가에 대한 이득의 감소가 선형에 가깝게 나타나는 동시에 속도 섭동에 대해 매우 빠른 반응을 보이는 것을 확인하였다. 2개의 서로 다른 화염전달함수가 적용된 축소차수모델 계산은 예측 여부에 핵심적인 역할인 두 함수 사이의 위상 지연을 반영할 수 있게 수행되었고, 실제 실험 결과와 비교 및 검증되었다. 이러한 계산 및 실험 결과를 통해, 다단 연소 시스템에서는 1차와 2차 화염의 서로 다른 동역학이 자발 불안정을 개별적으로, 혹은 동시에 일으킬 수 있음을 확인하였다. 추가로, 2차 화염에 의해 발생하는 불안정은 높은 차수의 연소기 음향 모드에 결합하는 특징을 보여, 앞선 연구 결과와 잘 일치하는 것으로 판명되었다. 연소기 내부에서 비대칭 형태로 나타난 각 화염의 상호작용은, 1차와 3차 길이방향 고유 주파수를 포함하여, 해당 주파수들의 상호변조왜곡성분들로 구성된 복잡한 형태의 불안정을 일으켰다. 여기서 2차 화염은 3차 모드 불안정 거동에 해당하는 제트 병합으로 인해 화염면의 소염 뿐만 아니라, 1차 화염의 불안정에 의해 발생하는 횡단류 섭동에 결합해 횡방향 거동이 동시에 나타나는, 매우 복잡하며 불규칙적인 화염 거동을 보였다. 추가로, 2차 화염의 열 방출 섭동의 진폭은 횡단류 섭동의 방향에 크게 영향을 받는 것을 확인하였다. 본 연구에서는 순차적으로 수행된 실험을 통해 특이하고 주목할 만한 현상들을 관찰하였다. 그 결과, 이전에 알려지지 않았던 축방향 다단 연소가 적용된 연소기에서 발생하는 다양한 연소불안정 현상을 처음으로 확인하였으며, 두 화염 사이에서 발생하는 복잡한 열음향학적 상호작용의 근본적인 메커니즘을 명료하게 설명하였다. 본 논문에서 기술한 실험 및 분석적 발견은 향후 다단 연소 시스템이 적용된 가스터빈의 개발 및 운용 과정에서 필수적인 토대이자 효과적인 도구로 활용될 것으로 예상된다.

서지기타정보

서지기타정보
청구기호 {DAE 25013
형태사항 xiii, 128 p. : 삽화 ; 30 cm
언어 영어
일반주기 저자명의 한글표기: 최용석
지도교수의 영문표기: Kim, Kyu Tae
지도교수의 한글표기: 김규태
학위논문 학위논문(박사) - 한국과학기술원 : 항공우주공학과,
서지주기 References: p. 110-128
주제 Combustion dynamics
Gas turbine
Axial-fuel staging
Multinozzle array
Hydrogen
Flame transfer function
Reduced-order modeling
연소불안정
가스터빈
다단 연소 기법
다중노즐
수소
화염전달함수
축소차수모델링
QR CODE

책소개

전체보기

목차

전체보기

이 주제의 인기대출도서

Evolution of global mean surface temperature (GMST) over the period of instrumental observations [2.

Total primary energy supply by energy carrier group, 2020-2050 under the 1.5'C Scenario [3].

The relation between hydrogen power generation technologies and the hydrogen network [7].

Chemical properties of fuel candidates for gas turbine [1이.

(a) Cross-sectional view of modern heavy-duty gas turbine engine (Simens SGT-300 MD [12]) consisting ofthree representative parts: compressor, combustor, and turbine, (b) simplified diagram of gas turbine engine, and (c) ideal Brayton cycle described in terms of pressure and volume of the gas

Comparison of two combustor types: (left) a traditional diffusion-based combustor, (right) a lean-premixed combustor.

Gas turbine combustors utilizing various combustion methodologies: (a,b) stratified recir- culation zones [16, 17], (c, d, and e) axial and radial staged combustion [18, 19].

Illustration of material enhancements with firing temperature capability of industrial gas turbines (TBC - thermal barrier coatings, CMC - ceramic matrix composites) 46].

(a) 3D real-engine-based model [59 illustrating the concept of axial fuel staging combus- tion, (b) adiabatic flame temperature dependence ofNOx emissions [60], (c) expected enhancement in turndown ratio.

(a) Retrofiting methods for exsisting gas turbine combustor provided by GE Vernova [59], (b) A schematic of the CPSC system of the Ansaldo Energia's GT36 model [58], and (c) combustion chamber with sequential two-stage combustion designed and manufactured by Kawasaki Heavy Industries 62].

Before-and-after images of the injector components from gas turbines and liquid rockets: (a) gas turbine combustor injector [85], (b) liquid rocket injector 84|.

(a, c) Hypothetical dependence of the acoustic driving, H(A), and damping, D(A), as a function of the instability amplitude A, and (b, d) conceptual plots for supercritical and subcritical bifurcations, respectively. Graphs are enhanced and redrawn from the original by [85].

Types of combustion instabilities: (a) longitudinal direction, (b) radial direction, (c) intrinsic thermoacoustic instability, and (b) transverse instabilites [72, 93].

Three physical mechanisms: pathways summarizing the causality leading to heat release oscillations through (a)velocity, (b) equivalence ratio, and (c) pressure [146].

(a) Classical schematic of the transverse jet and relevant vortical structures [172], (b) reacting jet-in-crossflow of lean-premixed hydrogen/air mixture [175].

(a) Measured temperature field of reacting fuel jet via Raman scattering method [184], (b) flame index of fuel reacting jet using a DNS calculation [189], and (c) specified autoignition regime of fuel reacting jet, using OH PLIF measurement 190].

(a) OH* chemilurninescence images for equivalence ratio variations at a constant jet-to- crossflow momentum ratio [175], and (b) comparison between non-reacting (closed symbols) and reacting (open symbols) JICF jet trajectories [203]. Images were modified and re-scaled to improve resolution.

(a) Measured NOx emissions as a function ofincreased temperature through the second- stage combustion (AT) [186], and (b) dependence of reacting jet NOx contribution on the equivalence ratio of the secondary jet, with various parameters presented as overlaid contours 227|.

3D CAD model of the complete experimental facility, featuring an entire laboratory-scale axial fuel staging capable model gas turbine combustor, detailed optical measurement setups regarding laser devices and photomiltipliers, and emission measurement setups.

(a) Laboratory-scale axial fuel staging combustion test facility. (b) Top view of primary injector configuration with constituent injector diameter of 6.5 土 0.03 mm. The vertical position of the pulsed UV laser sheet is marked on the nozzle array (light purple). (c) Cross-sectional view of the secondary combustor and spatial arrangement of nine circular injectors with an inner diameter of 6.5 士 0.

(a) OH PLIF and high-speed OH chemilurninescence measurement setup for secondary jet-in-crossflow section. Top view of 112.2 X 112.2 mm2 square combustor section is depicted together with planar sheet optics and laser beam path. DM = dichroic mirror. (b, c) Photographs of installed primary secondary injectors.

(a) Laboratory-scale axial fuel staging combustion test facility for self-excited instabilit measurement. Experimental setup for OH PLIF and high-speed OH* imaging measurements is showl schematically. (b, c) Modified inlet sections for transfer function measurements of upstream primar, flame and secondary jet-in-crossflow flame, respectively. Positions of two loudspeakers and hot wire probes are i

The usual chemiluninescence spectrum left after removing the effects of thermal radiation and widespread background emissions [267].

(a) Direct metal laser melting-based additive manufactured convergent nozzle. (b Schematic of constant temperature anemometer.

(a) The relationship between the OH-LIF signal and the excitation wavelength at a tem- perature of 2236 K and a pressure of1 atm, under the premise of vibrational non-equilibrium [269]. (b) experimentally measured OH fluorescence intensity distribution as a function of excitation wavelength. The peak is detected at 283.087 nm, where the laser operation is tuned to obtain high-quality image data.

A comprehensive illustration summarizing various imaging measurement techniques and their associated devices.

The schematic representation of the mathematical procedure for (a) spatial fast Fourier transform and (b) dynamic mode decomposition [27이.

The list of variables, instrument models, and corresponding uncertainties.

The list of measurements and utilized experimental device.

Projection of 3D structure of combustor into simplified 1D acoustic network model.

COMSOL calculation results for a axially-staged combustor model.

(a) Test matrix in terms of secondary equivalence ratio and secondary thermal power combinations. Filled circles denote sixteen fuel staging conditions, b to q, and baseline condition without secondary injection, test point a, as listed in Table 1 Total thermal power and primary equivalence ratio are constant at 80 kW and 0.60 for all test points, irrespective of fuel staging conditions. (b, c) Li

Test conditions for the measurements of combustion instabilities under stagednon-staged conditions.

Time-averaged line-of-sight integrated OH chemiluninescence emission intensities for pri- mary and secondary flames measured under sixteen test points (cases b to q) as well as baseline conditior (case a; primary injector only).

Instantaneous OH PLIFimages are aligned in the same visualizing format for 60X 60mm2 local interrogation regions, as marked with a white square in Fig. 3.2

Self-excited instability measurement data under baseline condition. (a) Measured instability frequency plotted against combustor length (black filled circles). Curved line and related gray shading indicate mean and one standard deviation of longitudinal-mode eigenfrequencies obtained from three dimensional Helmholtz simulations to account for temperature variations at exit boundary betweer 700 K a

(a, b) Phase-synchronized line-of-sight-integrated OH chemiluminescence images (inten- sified CCD camera measurement) and planar OH fluorescence images during a period of limit cycle oscillations at 443 Hz, corresponding to six consecutive phase angles illustrated together. Flow direction is from top to bottom in positive X direction. Whiteand blue dashed lines in (a) indicate local interro- gatio

(a) Normalized OH chemilurninescence emission intensity corresponding to L3 mode (443 Hz), with cutoff amplitude beingset to 10% of maximum amplitude (left). (b) Phase difference between acoustic pressure and heat release rate fluctuations obtained from high-speed OH chemiluminescence measurements (right).

For all test conditions (b- q), measured instability frequencies are plotted against combustor length using black empty circles, with curved lines representing the mean values of longitudnal-mode eigenfrequencies derived from three-dimensional Helmholtz simulations. Additionally, the right panel illustrates the probability of occurrence for each longitudinal mode.

Illustration ofisocontour stability maps in the Pth,sec, Osec> domain for constant combustoi lengths of 1300 (top), 1600 (middle), and 1800 mm (bottom row), expressed in terms of (a, d, g normalized combustor pressure amplitude at primary flame position, (b, e, h) normalized amplitude of primary flame's global heat release rate fluctuations, and (c, f, i) normalized amplitude of secondary flame's

(a, c) Time traces of acoustic pressure and global heat release rate fluctuations at primary (solid line) and secondary combustion zones (dashed line), with time normalizedbyfundamental period T = 2.227 and 1.653ms, for Cases A and B, respectively. Red circles in lower subpanels denote twelve consecutive phase angles used for phase-synchronized imaging measurements illustrated in Figs. 3.10and 3.1

Case A: self-excited dynamics of two axially-staged lean-premixed flames. (a, c) Phase synchronized OH chemiluminescence images for primary and secondary flames during a period oflimit cycle oscillations at 449 Hz. (b, d) Corresponding single-shot OH PLIF images for 60 X 60mm2 loca interrogation region marked in Figs. 3.10a and 3.10c for primary and secondary flames, respectively Flow directions a

Case B: self-excited dynamics of two axially-staged lean-premixed flames. (a, c) Phase synchronized OH chemilunninescence images for primary and secondary flames during a period of limi cycle oscillations at 605 Hz. (b, d) Corresponding single-shot OH PLIF images for 60 X 60mm2 loca interrogation region marked in Figs. 3.11a and 3.11c for primary and secondary flames, respectively Flow directions

Spectrum colormaps of (a) the dynamic pressure and heat release rate fluctuations for (b) the primary and (c) secondary flames. Fast Fourier Transform (FFT) results are shown as a function of frequency for five different H2 mole fraction conditions.

Amplitude-dependent relationships between dynamic pressure and the primary flame's heat release rate fluctuations at (a) 450 Hz and (b) 600 Hz. Data points are color-coded by hydrogen mole fraction for various conditions.

Time traces of acoustic pressure and heat release rate fluctuations, normalized by the fundamental period (T = 1/563 s), for (a) the primary and (b) secondary combustion zones at Case A (marked with a red circle in Fig. 3.13). All oscillating properties are bandpass-filtered at 563 Hz. In (b), velocity fluctuations at the secondary inlet plenum are shown alongside heat release rate fluctuations. R

Time traces of acoustic pressure and heat release rate fluctuations, normalized by the fundamental period (T = 1/565 s), for (a) the primary and (b) secondary combustion zones at Case B (marked with a red circle in Fig. 3.13). All oscillating properties are bandpass-filtered at 565 Hz. In (b), velocity fluctuations at the secondary inlet plenum are shown alongside heat release rate fluctuations. R

Left/right columns: Cases A and B. (a, b) Sequence of phase-resolved OH* chemilumi- nescence images for the primary and secondary flame during a period oflimit cycle oscillations.

(a) Illustration of test matrix, representing self-excited dynamics measurements with and without axial fuel staging (SED: black filled circles) and flame transfer function measurements for pri- mary and secondary flames (FTFpri and FTFsec: green open circles and red open circles, respectively) Different symbols are used to facilitate interpretation. (b, c) Heatmaps showing corresponding variation

Test conditions for self-excited dynamics (SED) and flame transfer function (FTF) measure- ments.

(a, b, c) Measured flame transfer functions for first-stage primary flame (test points a, c) and second-stage transverse jet flame plotted as a function of forcing frequency. Transfer functio measurement results of transverse jet flames are concerned with two different situations: primary equiv alence ratio is kept constant at 0.75 (test points f.i, l) and secondary equivalence ratio is kept const

(a) Low-order model prediction versus direct measurements of self-induced instability for three test points a, b, and C, with no axial staging, plotted as a function of effective combustoi length. Normalized growth rate is used to evaluate thermoacoustic stability, with positive and negative values corresponding to unstable and stable conditions. Gray shading signifies one standard deviatior assoc

(a, b, c) Heatmaps showing initial growth rate with respect to discrete variations of time delay (from -ㅠ to 十ㅠ in 0.17 radian increments) and effective combustor length (from 1200 to 1800 mm in 10 mm increments). (d) Phase difference ofinlet velocity disturbances between primary and secondary injectors; direct measurement using two hot wire probes under fuel-staged condition (test point l). (e) C

(a, b, c) Heatmaps showing initial growth rate with respect to discrete variations of time delay (from -ㅠ to 十ㅠ in 0.17 radian increments) and effective combustor length (from 1200 to 1800 mm in 10 mm increments) for case k.

(a) Test matrix encompassing nine conditions for self-excited dynamics measurement (re- plotting the matrix at 4.1a). (b, c) Amplitudes of heat release rate fluctuations for the primary and secondary flames in terms of the equivalence ratio (o) and frequency (f). These results represent data measured across the entire combustor length.

Normalized amplitude spectra of acoustic pressure, inlet velocity, and global heat release rate fluctuations, measured under test condition l; self-excited instability data measured under constant combustor length of 1300 mm. Dynamic pressure transducer positioned 500 mm from primary injector face was used to calculate pressure amplitude spectrum representing second-stage reaction zone; see Fig. 2

Phase-resolved instantaneous OH PLIF images of primary and secondary flames durin a single period of limit cycle oscillations at 312 Hz (top row) and 570 Hz (middle and bottom rows) Related spectral data are presented in Fig. 4.7 (test point l). 0o and 1800 phase angles are defined a minimum and maximum OH* emission points, respectively. Imaging field of view is indicated by blue dashed lines in F

(a, b) Two-dimensional OH* chemiluminescence amplitude spectra of primary flame en- semble (312 Hz) and secondary jet flame (258, 312, and 570 Hz), with cutoff amplitude set to 10 % of maximum amplitude. (c, d) Phase difference between local acoustic pressure and heat release rate fluctuations for primary and secondary reaction zones. Representative phase angles are annotated near isocontour lines

(a) Normalized amplitude spectra of heat release rate fluctuations for primary (top) and secondary (bottom) flames under conditions identical to those at case 1 (1300 mm). (b) Variations 0 the integrated OH* chemiluminescence intensity during a limit cycle period, including deviations, for both flames. Thedominant limit cycle frequencies of312 Hz and 570 Hz correspond to the peaks in each spectrum

(a) Temporal evolution ofthe normalized integrated OH* intensity ofthe secondary trans- verse jetflame over 0.03 seconds, with the local maxima highlighted. (b) Probability distribution function (PDF) ofthe local maxima of heat release from the entire measurement period (two seconds), derived from over 600 instantaneous images. (c) Ensemble-averaged OH* intensity images of the secondary flame corr