Engine Testing of a Natural Gas-Fired, Low-NOx, Variable Geometry Gas Turbine Combustor for a Small Gas Turbine

Author(s):  
Shigeru Hayashi ◽  
Hideshi Yamada ◽  
Kazuo Shimodaira

The development of a variable geometry lean-premixed combustor is in progress at NAL. Engine testing has been cooducted by using a natural gas-fueled 210-kW gas turbine to demonstrate the capability of ultra-low NOx emissions over a wide range of eogine operation. This paper describes the effort of engine testing of the combustor to achieve NOx emissions of the 10-ppm level. Fuel was staged to the non-premixed pilot and premixed main burners. A butterfly valve air splitting system was employed to maintain both low NOx emissions and high efficieocy over a wide operating range of the engine. The engioe was operated in the lean-premixed, low NOx emissions mode from idle to full power. Over the whole operating conditions from idle to full power, NOx emissions were reduced to levels less than 25 ppm (15% O2 dry). The NOx emissions level for a nearly constant combustion efficiency decreased with increasing power or turbine inlet temperature. At operating conditions of 90% to full power, NOx emissions levels of 12 to 8 ppm (15% O2 dry) were measured with combustion efficiencies of 99.7 to 99.1%.

Author(s):  
K. O. Smith ◽  
A. C. Holsapple ◽  
H. K. Mak ◽  
L. Watkins

The experimental results from the rig testing of an ultra-low NOx, natural gas-fired combustor for an 800 to 1000 kw gas turbine are presented. The combustor employed lean-premixed combustion to reduce NOx emissions and variable geometry to extend the range over which low emissions were obtained. Testing was conducted using natural gas and methanol. Testing at combustor pressures up to 6 atmospheres showed that ultra-low NOx emissions could be achieved from full load down to approximately 70% load through the combination of lean-premixed combustion and variable primary zone airflow.


Author(s):  
Elliot Sullivan-Lewis ◽  
Vincent McDonell

Ground based gas turbines are responsible for generating a significant amount of electric power as well as providing mechanical power for a variety of applications. This is due to their high efficiency, high power density, high reliability, and ability to operate on a wide range of fuels. Due to increasingly stringent air quality requirements, stationary power gas turbines have moved to lean-premixed operation. Lean-premixed operation maintains low combustion temperatures for a given turbine inlet temperature, resulting in low NOx emissions while minimizing emissions of CO and hydrocarbons. In addition, to increase overall cycle efficiency, engines are being operated at higher pressure ratios and/or higher combustor inlet temperatures. Increasing combustor inlet temperatures and pressures in combination with lean-premixed operation leads to increased reactivity of the fuel/air mixture, leading to increased risk of potentially damaging flashback. Curtailing flashback on engines operated on hydrocarbon fuels requires care in design of the premixer. Curtailing flashback becomes more challenging when fuels with reactive components such as hydrogen are considered. Such fuels are gaining interest because they can be generated from both conventional and renewable sources and can be blended with natural gas as a means for storage of renewably generated hydrogen. The two main approaches for coping with flashback are either to design a combustor that is resistant to flashback, or to design one that will not anchor a flame if a flashback occurs. An experiment was constructed to determine the flameholding tendencies of various fuels on typical features found in premixer passage ways (spokes, steps, etc.) at conditions representative of a gas turbine premixer passage way. In the present work tests were conducted for natural gas and hydrogen between 3 and 9 atm, between 530 K and 650K, and free stream velocities from 40 to 100 m/s. Features considered in the present study include a spoke in the center of the channel and a step at the wall. The results are used in conjunction with existing blowoff correlations to evaluate flameholding propensity of these physical features over the range of conditions studied. The results illustrate that correlations that collapse data obtained at atmospheric pressure do not capture trends observed for spoke and wall step features at elevated pressure conditions. Also, a notable fuel compositional effect is observed.


Author(s):  
Stefan Bauer ◽  
Balbina Hampel ◽  
Thomas Sattelmayer

Abstract Vortex generators are known to be effective in augmenting the mixing of fuel jets with air. The configuration investigated in this study is a tubular air passage with fuel injection from one single orifice placed in the side wall. In the range of typical gas turbine combustor inlet temperatures, the performance vortex generator premixers (VGPs) have already been investigated for natural gas as well as for blends of natural gas and hydrogen. However, for highly reactive fuels, the application of VGPs in recuperated gas turbines is particularly challenging because the high combustor inlet temperature leads to potential risk with regard to premature self-ignition and flame flashback. As the current knowledge does not cover the temperature range far above the self-ignition temperature, an experimental investigation of the operational limits of VGPs is currently being conducted at the Thermodynamics Institute of the Technical University of Munich, Garching, Germany, which is particularly focused on reactive fuels and the thermodynamic conditions present in recuperated gas turbines with pressure ratios of 4–5. For the study presented in this paper, an atmospheric combustion VGP test rig has been designed, which facilitates investigations in a wide range of operating conditions in order to comply with the situation in recuperated microgas turbines (MGT), namely, global equivalence ratios between 0.2 and 0.7, air preheating temperatures between 288 K and 1100 K, and air bulk flow rates between 6 and 16 g/s. Both the entire mixing zone in the VGP and the primary combustion zone of the test rig are optically accessible. High-speed OH* chemiluminescence imaging is used for the detection of the flashback and blow-off limits of the investigated VGPs. Flashback and blow-off limits of hydrogen in a wide temperature range covering the autoignition regime are presented, addressing the influences of equivalence ratio, air preheating temperature, and momentum ratio between air and hydrogen on the operational limits in terms of bulk flow velocity. It is shown that flashback and blow-off limits are increasingly influenced by autoignition in the ultrahigh temperature regime.


Author(s):  
K. O. Smith ◽  
G. W. Wade ◽  
M. H. Samii ◽  
H. K. Mak

The initial development of a lean-premixed, natural gas-fired combustor for a 200 kw gas turbine is described. The development effort included both rig testing and on-engine testing of the combustor. The combustor demonstrated an ultra-low NOx emissions capability in both test environments.


Author(s):  
Elliot Sullivan-Lewis ◽  
Vince McDonell

Ground-based gas turbines are responsible for generating a significant amount of electric power as well as providing mechanical power for a variety of applications. This is due to their high efficiency, high power density, high reliability, and ability to operate on a wide range of fuels. Due to increasingly stringent air quality requirements, stationary power gas turbines have moved to lean-premixed operation. Lean-premixed operation maintains low combustion temperatures for a given turbine inlet temperature, resulting in low NOx emissions while minimizing emissions of CO and hydrocarbons. In addition, to increase overall cycle efficiency, engines are being operated at higher pressure ratios and/or higher combustor inlet temperatures. Increasing combustor inlet temperatures and pressures in combination with lean-premixed operation leads to increased reactivity of the fuel/air mixture, leading to increased risk of potentially damaging flashback. Curtailing flashback on engines operated on hydrocarbon fuels requires care in design of the premixer. Curtailing flashback becomes more challenging when fuels with reactive components such as hydrogen are considered. Such fuels are gaining interest because they can be generated from both conventional and renewable sources and can be blended with natural gas as a means for storage of renewably generated hydrogen. The two main approaches for coping with flashback are either to design a combustor that is resistant to flashback, or to design one that will not anchor a flame if a flashback occurs. An experiment was constructed to determine the flameholding tendencies of various fuels on typical features found in premixer passage ways (spokes, steps, etc.) at conditions representative of a gas turbine premixer passage way. In the present work, tests were conducted for natural gas and hydrogen between 3 and 9 atm, between 530 K and 650 K, and free stream velocities from 40 to 100 m/s. Features considered in the present study include a spoke in the center of the channel and a step at the wall. The results are used in conjunction with existing blowoff correlations to evaluate flameholding propensity of these physical features over the range of conditions studied. The results illustrate that correlations that collapse data obtained at atmospheric pressure do not capture trends observed for spoke and wall step features at elevated pressure conditions. Also, a notable fuel compositional effect is observed.


Author(s):  
Jon P. McDonald ◽  
Arthur M. Mellor

Semi–empirical characteristic time models (CTMs) for NOx emissions index (EI) and lean blowoff are used in the design of an inlet condition matrix for measurement of NOxEI from a lean premixed combustor. Such models relate either NOxEI or the weak extinction limit to times representing relevant physical and chemical processes in the combustor. Lean premixed (LP) natural gas/air combustion is considered for the following conditions: inlet temperature, 300–800 K; combustor pressure, 1–30 atm; and equivalence ratio, 0.5–0.7. The NOx model is used to determine combinations of inlet conditions corresponding to greatest NOx sensitivity. A dependence of NOx emissions on pressure is included in the model. Emissions of oxides of nitrogen are found to he most sensitive to variations in inlet temperature and combustor pressure, in the 560–800 K and 20–30 atm ranges, respectively, while sensitivity to variations in equivalence ratio is substantial over the entire range considered. Thus it is found that operating conditions for high thermal efficiency in LP turbine combustors conflict with the goal of lowering NOx emissions, a result consistent with thermal NOx from conventional, diffusion flame combustors. A lean blowoff model is used to estimate the lowest equivalence ratio at which a flame can he held, as well as to determine whether a flame can be stabilised at the operating conditions suggested by the NOx sensitivity analysis. The results suggest a nominal lower limit on equivalence ratio of 0.4, and that a flame can be held for most of the combinations of inlet conditions suggested by the NOx sensitivity analysis. Autoignition of the fuel/air mixture is also considered in relation to the location and/or design of the premixing system. The current NOx CTM is applied to LP natural gas fired data from the literature. A model modification, thought to better represent the fluid mechanics relevant to LP NOx formation, is applied, and its implications discussed.


Author(s):  
K. O. Smith ◽  
A. Fahme

Three subscale, cylindrical combustors were rig tested on natural gas at typical industrial gas turbine operating conditions. The intent of the testing was to determine the effect of combustor liner cooling on NOx and CO emissions. In order of decreasing liner cooling, a metal louvre-cooled combustor, a metal effusion-cooled combustor, and a backside-cooled ceramic (CFCC) combustor were evaluated. The three combustors were tested using the same lean-premixed fuel injector. Testing showed that reduced liner cooling produced lower CO emissions as reaction quenching near the liner wall was reduced. A reduction in CO emissions allows a reoptimization of the combustor air flow distribution to yield lower NOx emissions.


Author(s):  
K. K. Botros ◽  
H. Golshan ◽  
D. Rogers ◽  
B. Sloof

Gas turbine (GT) engines employed in natural gas compressor stations operate in different modes depending on the power, turbine inlet temperature and shaft speeds. These modes apply different sequencing of bleed valve opening on the air compressor side of the engine. Improper selection of the GT and the driven centrifugal gas compressor operating conditions can lead to larger bleed losses due to wider bleed valve openings. The bleed loss inevitably manifests itself in the form of higher overall heat rate of the GT and greater engine emission. It is therefore imperative to determine and understand the engine and process conditions that drive the GT to operate in these different modes. The ultimate objective is to operate the engine away from the inefficient modes by adjusting the driven gas compressor parameters as well as the overall station operating conditions (i.e. load sharing, control set points, etc.). This paper describes a methodology to couple the operating conditions of the gas compressor to the modes of GT bleed valve opening (and the subsequent air bleed rates) leading to identification of the operating parameters for optimal performance (i.e., best overall efficiency and minimum CO2e emission). A predictive tool is developed to quantify the overall efficiency loss as a result of the different bleed opening modes, and map out the condition on the gas compressor characteristics. One year’s worth of operating data taken from two different compressor stations on TransCanada Pipelines’ Alberta system were used to demonstrate the methodology. The first station employs GE-LM1600 gas turbine driving a Cooper Rolls-RFBB-30 centrifugal compressor. The second station employs GE-LM-2500+ gas turbine driving NP PCL-800/N compressor. The analysis conclusively indicates that there are operating regions on the gas compressor maps where losses due to bleed valves are reduced and hence CO2 emissions are lowered, which presents an opportunity for operation optimization.


Author(s):  
Iarno Brunetti ◽  
Giovanni Riccio ◽  
Nicola Rossi ◽  
Alessandro Cappelletti ◽  
Lucia Bonelli ◽  
...  

The use of hydrogen as derived fuel for low emission gas turbine is a crucial issue of clean coal technology power plant based on IGCC (Integrated Gasification Combined Cycle) technology. Control of NOx emissions in gas turbines supplied by natural gas is effectively achieved by lean premixed combustion technology; conversely, its application to NOx emission reduction in high hydrogen content fuels is not a reliable practice yet. Since the hydrogen premixed flame is featured by considerably higher flame speed than natural gas, very high air velocity values are required to prevent flash-back phenomena, with obvious negative repercussions on combustor pressure drop. In this context, the characterization of hydrogen lean premixed combustion via experimental and modeling analysis has a special interest for the development of hydrogen low NOx combustors. This paper describes the experimental and numerical investigations carried-out on a lean premixed burner prototype supplied by methane-hydrogen mixture with an hydrogen content up to 100%. The experimental activities were performed with the aim to collect practical data about the effect of the hydrogen content in the fuel on combustion parameters as: air velocity flash-back limit, heat release distribution, NOx emissions. This preliminary data set represents the starting point for a more ambitious project which foresees the upgrading of the hydrogen gas turbine combustor installed by ENEL in Fusina (Italy). The same data will be used also for building a computational fluid dynamic (CFD) model usable for assisting the design of the upgraded combustor. Starting from an existing heavy-duty gas turbine burner, a burner prototype was designed by means of CFD modeling and hot-wire measurements. The geometry of the new premixer was defined in order to control turbulent phenomena that could promote the flame moving-back into the duct, to increase the premixer outlet velocity and to produce a stable central recirculation zone in front of the burner. The burner prototype was then investigated during a test campaign performed at the ENEL’s TAO test facility in Livorno (Italy) which allows combustion test at atmospheric pressure with application of optical diagnostic techniques. In-flame temperature profiles, pollutant emissions and OH* chemiluminescence were measured over a wide range of the main operating parameters for three fuels with different hydrogen content (0, 75% and 100% by vol.). Flame control on burner prototype fired by pure hydrogen was achieved by managing both the premixing degree and the air discharge velocity, affecting the NOx emissions and combustor pressure losses respectively. A CFD model of the above-mentioned combustion test rig was developed with the aim to validate the model prediction capabilities and to help the experimental data analysis. Detailed simulations, performed by a CFD 3-D RANS commercial code, were focused on air/fuel mixing process, temperature field, flame position and NOx emission estimation.


Author(s):  
Paul A. Dellenback

An alternative configuration for a regenerative gas turbine engine cycle is presented that yields higher cycle efficiencies than either simple or conventional regenerative cycles operating under the same conditions. The essence of the scheme is to preheat compressor discharge air with high temperature combustion gases before the latter are fully expanded across the turbine. The efficiency is improved because air enters the combustor at a higher temperature, and hence heat addition in the combustor occurs at a higher average temperature. The heat exchanger operating conditions are more demanding than for a conventional regeneration configuration, but well within the capability of modern heat exchangers. Models of cycle performance exhibit several percentage points of improvement relative to either simple cycles or conventional regeneration schemes. The peak efficiencies of the alternative regeneration configuration occur at optimum pressure ratios that are significantly lower than those required for the simple cycle. For example, at a turbine inlet temperature of 1300°C (2370°F), the alternative regeneration scheme results in cycle efficiencies of 50% for overall pressure ratios of 22, whereas simple cycles operating at the same temperature would yield efficiencies of only 43.8% at optimum pressure ratios of 50, which are not feasible with current compressor designs. Model calculations for a wide range of parameters are presented, as are comparisons with simple and conventional regeneration cycles.


Sign in / Sign up

Export Citation Format

Share Document