Flashback is a critical operability issue for lean premixed fuel injection strategies. This is particularly true for hydrogen and high hydrogen content fuels. As a result, efficient analytical tools to help predict the occurrence of flashback would be of great value to the hardware designer. In the present work, a RANS based CFD strategy is used in conjunction with analytical expressions for determining flashback likelihood. The approach uses local flow properties combined with 1 D models to establish a flashback propensity factor that can be used to screen designs. Peclet number based models are also considered. The methods are applied to one micromixer configuration for a variety flow rates, pressures, equivalence ratios, and temperatures. The micromixing injector is a small premixing cup with multiple fuel injection sites in close proximity to air inlets to produce rapid mixing between fuel and air, with an axial core flow that is intended to prevent ingress of flame into the injector. The results illustrate the effectiveness of the approach for screening for flashback for the premixer studied over the range of conditions studied. The analysis suggests that, for the current test hardware, the critical ingress pathway is via boundary layer propagation. Further, the approach successfully provides a qualitative indication of whether ingress can be expected for a given condition.

The present work extends previous efforts using “micro-mixing” fuel injectors operating on hydrogen fuel to elevated pressure and temperature and includes initial evaluation of a second injector concept. A micro-mixing fuel injector consists of multiple, small and closely spaced mixing cups, within which fuel and air mix rapidly at a small scale. The micro-mixing injection strategy offers inherent flexibility for the accommodation of staging, dilution, and fuel flexibility, and the manufacturing technology employed for building the cups affords great flexibility to address the conflicting demands of superior fuel-air mixing and flash-back avoidance. In the present work, both radial and axial flow micromixing concepts are investigated using experiments and computational fluid dynamics. The hydrogen/air reaction structure is captured using OH* chemiluminescence at 308 nm, recorded using a 16-bit thermoelectrically cooled ICCD with a UV sensitive phosphor. Instantaneous images are used to assess flashback tendencies at pressures up to 8 atm for reaction temperatures approaching 2000 K. Emissions are measured at the exit of the combustor liner using EPA certified methodologies. The results demonstrate that both concepts can produce low NOx emissions while remaining robust relative to flashback and lean blowout. The radial concepts offer superior emissions performance, while the axial concepts offer superior flashback tendencies. Based on the results obtained to date, the micro-mixing approach appears promising relative to achieving flashback free operation with low emissions at pressures up to 8 atm while maximizing scalability and fuel flexibility.

The consideration of hydrogen as a fuel for next generation low emissions gas turbines raises a number of challenges and potential benefits relative to the combustion system. The present work examines the use of a micro-mixing injection strategy for hydrogen as a means to achieve rapid mixing and inherent flexibility for accommodating various staging, dilution, and dual fuel requirements for future gas turbine engines. The work presented includes numerical and experimental results associated with the fuel-air mixing process in a representative injector configuration. Measured NOx emissions and fuel/air ratios at the exit of the mixer are shown along with visualization of the reactions generated. Detailed computational fluid dynamics (CFD) is used in parallel to elucidate the behavior of the flow inside and downstream of the injectors. Results are also presented for natural gas to provide a point of reference. The results illustrate a number of interesting features and characteristics of the hydrogen/air mixtures which are in dramatic contrast to the behavior of natural gas/air mixtures. Comparison of the measured and modeled mixing behavior illustrates a number of challenges associated with the selection of a robust modeling approach for hydrogen/air combustion. The results demonstrate that the use of micro-mixing fuel injection to achieve ultra low NOx emissions is very promising.

This paper discusses the development and testing of a full-scale micro-mixing lean-premix injector for hydrogen and syngas fuels that demonstrated ultra-low emissions and stable operation without flashback for high-hydrogen fuels at representative full-scale operating conditions. The injector was fabricated using Macrolamination technology, which is a process by which injectors are manufactured from bonded layers. The injector utilizes sixteen micro-mixing cups for effective and rapid mixing of fuel and air in a compact package. The full scale injector is rated at 1.3 MWth when operating on natural gas at 12.4 bar (180 psi) combustor pressure. The injector operated without flash back on fuel mixtures ranging from 100% natural gas to 100% hydrogen and emissions were shown to be insensitive to operating pressure. Ultra-low NOx emissions of 3 ppm were achieved at a flame temperature of 1750 K (2690 °F) using a fuel mixture containing 50% hydrogen and 50% natural gas by volume with 40% nitrogen dilution added to the fuel stream. NOx emissions of 1.5 ppm were demonstrated at a flame temperature over 1680 K (2564 °F) using the same fuel mixture with only 10% nitrogen dilution, and NOx emissions of 3.5 ppm were demonstrated at a flame temperature of 1730 K (2650 °F) with only 10% carbon dioxide dilution. Finally, using 100% hydrogen with 30% carbon dioxide dilution, 3.6 ppm NOx emissions were demonstrated at a flame temperature over 1600 K (2420 °F). Superior operability was achieved with the injector operating at temperatures below 1470 K (2186 °F) on a fuel mixture containing 87% hydrogen and 13% natural gas. The tests validated the micro-mixing fuel injector technology and the injectors show great promise for use in future gas turbine engines operating on hydrogen, syngas or other fuel mixtures of various compositions.

Participating in NASA’s Environmentally Responsible Aviation (ERA) Project, Parker Hannifin built and tested multipoint Lean Direct Injection (LDI) fuel injectors designed for NASA’s N+2 55:1 Overall Pressure-Ratio (OPR) gas turbine engine cycles. The injectors are based on Parker’s earlier three-zone injector (3ZI) which was conceived to enable practical implementation of multipoint LDI schemes in conventional aviation gas turbine engines. The new injectors offer significant aerodynamic design flexibility, excellent thermal performance, and scalability to various engine sizes. The injectors built for this project contain 15 injection points and incorporate staging to enable operation at low power conditions. Ignition and flame stability were demonstrated at ambient conditions with ignition air pressure drop as low as 0.3% and fuel-to-air ratio (FAR) as low as 0.011. Lean Blowout (LBO) occurred at FAR as low as 0.005 with air at 460 K and atmospheric pressure. A high pressure combustion testing campaign was conducted in the CE-5 test facility at NASA Glenn Research Center at pressures up to 250 psi and combustor exit temperatures up to 2,033 K (3,200 °F). The tests demonstrated estimated LTO cycle emissions that are about 30% of CAEP/6 for a reference 60,000 lbf thrust, 54.8-OPR engine. This paper presents some details of the injector design along with results from ignition, LBO and emissions testing.

No abstract

Measurements of the flow field around a low pressure turbine blade are presented. The purpose of the experimental study is to investigate the detailed transition and separation characteristics on low pressure turbine blades. The present study focuses on the effect of Reynolds number on the formation of separation bubbles on the turbine blade surface. Specific attention is paid to the flow within and around the region of separation. Experiments have been conducted using a turbine blade cascade in a tow tank and wind tunnel using DPIV and flow visualization, respectively. In the tow tank, a turning vane cascade model was constructed to validate the diagnostics for future wind tunnel tests. DPIV has been employed to measure the flow velocity near the surface of the blade with a wall-adaptive Lagrangian parcel tracking algorithm which enables the determination of velocities near surface boundaries. Since this technique is instantaneous, it allows investigation of unsteady flow phenomenon critical in understanding transient separation. Measurements are run for a Re range from 1 • 10 4 to 1 • 10 5 ; preliminary measurements are presented from 1 • 10 4 to 4 • 10 4. In the wind tunnel, a 6 blade cascade model was constructed with a variable exit angle for flow visualization and DPIV. Smoke wire visualization results are presented for Re 2.5 • 10 4 to 9 • 10 4. Nomenclature Re Reynolds number β Exit angle, deg Subscripts c Chord length SSL Suction surface length ∞ Freestream d Wire diameter

No abstract