Combustion Characteristics and NOX Emission of Hydrogen-Rich Fuel Gases at Gas Turbine Relevant Conditions

Author(s):  
Y.-C. Lin ◽  
S. Daniele ◽  
P. Jansohn ◽  
K. Boulouchos

In this paper, characteristics of turbulent combustion and NOx emission for high hydrogen-content fuel gases (H2 > 70 vol. %; “hydrogen-rich”) are addressed. An experimental investigation is performed in a perfectly-premixed axial-dump combustor under gas turbine relevant conditions. Fundamental features of turbulent combustion for these mixtures are evaluated based on OH-PLIF diagnostics. On the other hand, NOx emissions are measured with an exhaust gas sampling probe positioned downstream the combustor outlet. Compared to syngas mixtures (H2 + CO), the operational limits for hydrogen-rich fuel gases are found to occur at even leaner conditions concerning flashback phenomena. With respect to effects of operating pressure, a strongly reduced operational envelope is observed at elevated pressure. Only with decreasing the preheat temperature a viable approach to further extend the operational range is seen. Evaluation of the averaged turbulent flame shape shows that the profile of the flame front is generally approaching that of an ideal cone. Thus a simplified approach for estimating the turbulent flame speed via the location of the flame tip alone can be applied. The level of NOx emission for the hydrogen-rich fuel mixtures is generally above that of syngas mixtures, which exhibit already higher NOx emission values than natural gas. Distinct chemical kinetic features are found specifically at elevated pressure. While the pressure effects are weak for syngas, a non-monotonic behavior is observed for the hydrogen-rich fuels. Reaction path analysis is performed to complement and provide more insight to the findings from the measurements. From chemical kinetic calculations a distinct shift in NOx formation pathways (thermal NOx vs. NOx through N2O/NNH reaction channels) can be observed for the different fuel mixtures at different pressure levels.


Author(s):  
Yu-Chun Lin ◽  
Salvatore Daniele ◽  
Peter Jansohn ◽  
Konstantinos Boulouchos

The turbulent flame speed (ST) is proposed to be an indicator of the flashback propensity for hydrogen-rich fuel gases at gas turbine relevant conditions. Flashback is an inevitable issue to be concerned about when introducing fuel gases containing high hydrogen content to gas turbine engines, which are conventionally fueled with natural gas. These hydrogen-containing fuel gases are present in the process of the integrated gasification combined cycle (IGCC), with and without precombustion carbon capture, and both syngas (H2 + CO) and hydrogen with various degrees of inert dilution fall in this category. Thus, a greater understanding of the flashback phenomenon for these mixtures is necessary in order to evolve the IGCC concept (either with or without carbon capture) into a promising candidate for clean power generation. Compared to syngas, the hydrogen-rich fuel mixtures exhibit an even narrower operational envelope between the occurrence of lean blow out and flashback. When flashback occurs, the flame propagation is found to occur exclusively in the boundary layer of the pipe supplying the premixed fuel/air mixture to the combustor. This finding is based on the experimental investigation of turbulent lean-premixed nonswirled confined jet flames for three fuel mixtures with H2 > 70 vol. %. Measurements were performed up to 10 bar at a fixed bulk velocity at the combustor inlet (u0 = 40 m/s) and preheat temperature (T0 = 623 K). Flame front characteristics were retrieved via planar laser-induced fluorescence of the hydroxyl radical (OH-PLIF) diagnostics and the turbulent flame speed (ST) was derived, accordingly, from the perspective of a global consumption rate. Concerning the flashback limit, the operational range of the hydrogen-rich mixtures is found to be well represented by the velocity gradients prescribed by the flame (gc) and the flow (gf), respectively. The former (gc) is determined as ST/(Le × δL0), where Le is the Lewis number and δL0 is the calculated thermal thickness of the one-dimensional laminar flame. The latter (gf) is predicted by the Blasius correlation for fully developed turbulent pipe flow and it indicates the capability with which the flow can counteract the opposed flame propagation. Our results show that the equivalence ratios at which the two velocity gradients reach similar levels correspond well to the flashback limits observed at various pressures. The methodology is also found to be capable of predicting the aforementioned difference in the operational range between syngas and hydrogen-rich mixtures.



Author(s):  
Xunwei Liu ◽  
Weiwei Shao ◽  
Yong Tian ◽  
Yan Liu ◽  
Bin Yu ◽  
...  

For high-hydrogen-content fuel, the Micromix Combustion Technology has been developed as a potential low NOx emission solution for gas turbine combustors, especially for advanced gas turbines with high turbine inlet temperature. Compared with conventional lean premixed flames, multiple distributed slim and micro flames could lead to a lower NOx emission performance for shortening residence time of high temperature flue gas and generally a more uniform temperature distribution. This work aims at micromix flame characteristics of a model burner fueled with hydrogen blending with methane under atmosphere pressure conditions. The model burner assembly was designed to have six concentrically millimeter-sized premixed units around a same unit centrally. Numerical and experimental studies were conducted on mixing performance, flame stability, flame structure and CO/NOx emissions of the model burner. OH radical distribution by OH-PLIF and OH chemiluminescence (OH*) imaging were employed to analyze the turbulence-reaction interactions and characters of the reaction zone at the burner exit. Micromix flames fueled with five different hydrogen content H2-CH4 (60/40, 50/50, 40/60, 30/70, 0/100 Vol.%) were investigated, along with the effects of equivalence ratio and heat load. Results indicated that low NOx emissions of less than 10 ppm (@15% O2) below the exhaust temperature of 1920 K were obtained for all the different fuels. Combustion oscillation didn’t occur for all the conditions. It was found that at a constant flame temperature, the higher the hydrogen content of the fuel, the higher the turbulent flame speed and the weaker the flame lift effect. Combustion noise and NOx emissions also increase with increasing hydrogen content. The OH/OH* signal distribution indicated that a pure methane micromix flame showed a lifted and weaken distributed feature.



Author(s):  
Anibal Morones ◽  
Mattias A. Turner ◽  
Victor León ◽  
Kyle Ruehle ◽  
Eric L. Petersen

Abstract Turbulent combustion is a very active and challenging research topic of direct interest to the design and operation of gas turbine engines. A spherically expanding flame immersed in a turbulent field is one way to gain fundamental insight on the effect of turbulence on combustion. This kind of experiment is often conducted inside a fan-stirred flame bomb, preferably at conditions of high pressure, high temperature, and intense turbulence. A new fan-stirred flame bomb was designed and built to provide a device for conducting fundamental turbulent flame measurements at conditions of interest to gas turbine engines. A literature review on existing systems was used as guidance in the design of the turbulence-generation elements in the present rig. A few options of impellers were explored. The flow field produced by the chosen impeller was measured with Laser Doppler Velocimetry (LDV). A detailed exposition of the vessel engineering and construction are presented, including current activities that will extend the use of the facility for heated experiments up to at least 400 K. Before turbulent experiments were attempted, a validation of the rig accuracy and pressure worthiness was made. Finally, a demonstration of the new apparatus was made by testing a lean mixture of syngas. The experiment matrix using hydrogen and H2/CO mixtures included three levels of pressure (1, 5, and, 10 bar) and three levels of turbulence fluctuation rms (1.4, 2.8, and 5.5 m/s). Data based on the high-speed schlieren diagnostic are presented.



Author(s):  
Y.-C. Lin ◽  
S. Daniele ◽  
P. Jansohn ◽  
K. Boulouchos

The turbulent flame speed (ST) is proposed to be an indicator of flashback propensity for hydrogen-rich fuel gases at gas turbine relevant conditions. Flashback is an inevitable issue to be concerned about when introducing fuel gases containing high hydrogen content to gas turbine engines, which are conventionally fueled with natural gas. These hydrogen-containing fuel gases are present in the process of the integrated gasification combined cycle (IGCC) with and without pre-combustion carbon capture, and both syngas (H2 + CO) and hydrogen with various degree of inert dilution fall in this category. More understanding on the flashback phenomenon for these mixtures is thus necessary in order to evolve the IGCC concept (either with or without carbon capture) into a promising candidate for clean power generation. Compared to syngas, the hydrogen-rich fuel mixtures exhibit an even narrower operational envelope between the occurrence of lean blow out and flashback. When flashback occurs, the flame propagation is found to happen exclusively in the boundary layer of the pipe supplying the premixed fuel/air mixture to the combustor. This finding is based on the experimental investigation of turbulent, lean-premixed, non-swirled, confined jet flames for three fuel mixtures with H2 > 70 Vol. %. Measurements were performed up to 10 bar at fixed bulk velocity at the combustor inlet (u0 = 40 m/s) and preheat temperature (T0 = 623 K). Flame front characteristics were retrieved via OH-PLIF diagnostics, and turbulent flame speed (ST) was derived accordingly from the perspective of a global consumption rate. Concerning the flashback limit, the operational range of the hydrogen-rich mixtures is found to be well represented by the velocity gradients prescribed respectively by the flame (gc) and the flow (gf). The former (gc) is determined as ST/(Le × δL0), where Le is the Lewis number and δL0 is the calculated thermal thickness of the one-dimensional laminar flame. The latter (gf) is predicted by the Blasius correlation for fully developed turbulent pipe flow, and it indicates the capability with which the flow can counteract the opposed flame propagation. Our results show that the equivalence ratios at which the two velocity gradients reach similar levels correspond well to the flashback limits observed at various pressures. The methodology is also found to be capable of predicting the aforementioned difference in the operational range between syngas and hydrogen-rich mixtures.



Author(s):  
S. Daniele ◽  
P. Jansohn ◽  
K. Boulouchos

This paper focuses on the description of the turbulent flame speed, at gas turbine like conditions, for different syngas mixtures, selected in order to simulate syngas compositions typically derived from gasification of coal, oil, biomass, and used for power generation in integrated gasification combined cycle (IGCC) processes. In this paper the turbulent flame speed is reported as global consumption rate and calculated based on a mass continuity approach applied to the combustor inlet area and the flame front surface, which was detected experimentally. Flame front detection was done by means of planar laser induced fluorescence technique taking OH radicals as seeding dyes. An in-house developed flame front detection software tool has been further improved and utilized in this work in order to better fit ultra-lean H2-rich flames. Experiments were carried out in a High Pressure Test Rig for operating pressures up to 15 bar. Data provided in this paper will focus on a pressure level of 5 bar, adiabatic flame temperatures up to 1900 K, inlet velocities from 40 to 80 m/s, and inlet temperature of 672 and 772 K. As expected, the results highlight the strongly elevated values of turbulent flame speed for high hydrogen containing fuel gas mixtures. Compared with flame speed data for pure CH4 the ratio (STSyn/STCH4) takes up values of 7 to 8. In absolute terms values go up even beyond 10 m/s. With increased H2 content in the mixture the burning velocity raises, due to the faster chemical kinetics characteristic of this compound and due to physical properties of H2 (Le<1) which enhance flame front corrugation (i.e. flame front surface). Inlet velocity and pressure variations showed to have weak effect on the average flame front position whereas this last parameter is strongly affected by the mixture composition, the equivalence ratio and inlet temperature.



Author(s):  
Vladimir Zimont ◽  
Wolfgang Polifke ◽  
Marco Bettelini ◽  
Wolfgang Weisenstein

Theoretical background, details of implementation and validation results of a computational model for turbulent premixed gaseous combustion at high turbulent Reynolds numbers are presented. The model describes the combustion process in terms of a single transport equation for a progress variable; closure of the progress variable’s source term is based on a model for the turbulent flame speed. The latter is identified as a parameter of prime significance in premixed turbulent combustion and is determined from theoretical considerations and scaling arguments, taking into account physico-chemical properties of the combustible mixture and local turbulent parameters. Specifically, phenomena like thickening, wrinkling and straining of the flame front by the turbulent velocity field are considered, yielding a closed form expression for the turbulent flame speed that involves, e.g., speed, thickness and critical gradient of a laminar flame, local turbulent length scale and fluctuation intensity. This closure approach is very efficient and elegant, as it requires only one transport equation more than the non-reacting flow case, and there is no need for costly evaluation of chemical source terms or integration over probability density functions. The model was implemented in a finite-volume based computational fluid dynamics code and validated against detailed experimental data taken from a large scale atmospheric gas turbine burner test stand. The predictions of the model compare well with the available experimental results. It has been observed that the model is significantly more robust and computationally efficient than other combustion models. This attribute makes the model particularly interesting for applications to large 3D problems in complicated geometries.



Author(s):  
Matias Muller ◽  
Corbin Freeman ◽  
Peng Zhao ◽  
Haiwen Ge

The ignition mechanism of a lean premixed CHVair mixture by a hot turbulent jet issued from the pre-chamber combustion is investigated using 3D combustion CFD. The turbulent jet ignition experiments conducted in the rapid compression machine (RCM) at Michigan State University (MSU) were simulated. A full simulation was carried out first using RANS model for validation, the results of which were then taken as the boundary condition for the detailed simulations using both RANS and LES. To isolate the thermal and chemical kinetic effects from the hot jet, two different inlet conditions of the chamber were considered: inert case (including thermal effects only) and reactive case (accounting for both thermal and chemical kinetic effects). It is found that the chemical kinetic effects are important for the ignition in the main chamber. Comparison of OH and HRR (heat release rate) computed by RANS and LES shows that RANS predicts slightly faster combustion, which implies higher predicted turbulent flame speed. Correlations between vorticity, mixing field, and temperature field are observed, which indicate that the flow dynamics strongly influence the mixing process near the flame front, and consequently affect flame propagation.



Author(s):  
George Mallouppas ◽  
Graham Goldin ◽  
Yongzhe Zhang ◽  
Piyush Thakre ◽  
Jim Rogerson

Abstract Three Flamelet Generated Manifold reaction source term closure options and two different reactor types are examined with Large Eddy Simulation of an industrial gas turbine combustor operating at 3 bar. This work presents the results for the SGT-100 Dry Low Emission (DLE) gas turbine provided by Siemens Industrial Turbomachinery Ltd. The related experimental study was performed at the German Aerospace Centre, DLR, Stuttgart, Germany. The FGM model approximates the thermo-chemistry in a turbulent flame as that in a simple 0D constant pressure ignition reactors and 1D strained opposed-flow premixed reactors, parametrized by mixture fraction, progress variable, enthalpy and pressure. The first objective of this work is to compare the flame shape and position predicted by these two FGM reactor types. The Kinetic Rate (KR) model, studied in this work, uses the chemical rate from the FGM with assumed shapes, which are a Beta function for mixture fraction and delta functions for reaction progress variable and enthalpy. Another model investigated is the Turbulent Flame-Speed Closure (TFC) model with Zimont turbulent flame speed, which propagates premixed flame fronts at specified turbulent flame speeds. The Thickened Flame Model (TFM), which artificially thickens the flame to sufficiently resolve the internal flame structure on the computational grid, is also explored. Therefore, a second objective of this paper is to compare KR, TFC and TFM with the available experimental data.



Author(s):  
Alireza Kalantari ◽  
Vincent McDonell ◽  
Scott Samuelsen ◽  
Shahram Farhangi ◽  
Don Ayers

Lean premixed combustion is extensively used in gas turbine industry to reduce pollutant emissions. However, combustion stability still remains as a primary challenge associated with high hydrogen content fuels. Flashback is a crucial concern for designing gas turbine combustors in terms of operability limits. The current experimental study addresses the boundary layer flashback of hydrogen-air premixed jet flames at gas turbine premixer conditions (i.e. elevated pressure and temperature). Flashback propensity of a commercially available injector, originally designed for natural gas, is studied at different operating conditions and corresponding measurements are presented. The pressure dependence of flashback propensity is consistent with previous studies. The previously developed flashback model is successfully applied to the current data, verifying its utilization for various test conditions/setups. In addition, the model is used to predict flashback propensity of the injector at the actual engine preheat temperature. The injector is then modified to increase boundary layer flashback resistance and the corresponding data are collected at the same operating conditions. To avoid the boundary layer flashback, the mixture is leaned out in the near-wall region, where the flame can potentially propagate upstream. The comparison of gathered data shows a clear improvement in flashback resistance. This improvement is further elaborated by numerically studying fuel/air mixing characteristics for the two injectors.



2016 ◽  
Vol 30 (9) ◽  
pp. 7691-7703 ◽  
Author(s):  
Sheikh F. Ahmed ◽  
Jeffrey Santner ◽  
Frederick L. Dryer ◽  
Bihter Padak ◽  
Tanvir I. Farouk


Sign in / Sign up

Export Citation Format

Share Document