Marc LaViolette

Emissions From a Gas Turbine Sector Rig Operated With Synthetic Aviation and Biodiesel Fuel

Differences in exhaust emissions, smoke production, exhaust pattern factor, deposit buildup, and fuel nozzle spray characteristics for various blends of conventional commercial jet fuel (Jet A-1) with synthetic and biodiesel formulations were investigated. Three synthetic fuel formulations and four fatty acid methyl esters (FAMEs) were evaluated as such. The synthetic fuels were tested in both neat (100%) and 50% by volume blends with Jet A-1, while the FAME fuels were blended in 2% and 20% proportions. The combustion chamber sector rig, which houses a Rolls Royce T-56-A-15 combustion section, was utilized for emissions, deposits, and exhaust pattern factor evaluation. A combustion chamber exhaust plane traversing thermocouple rake was employed to generate two-dimensional temperature maps during operation. Following combustion testing, several combustion system components, including the combustion chamber, fuel nozzle, and igniter plug, were analyzed for relative levels of deposit buildup. A phase Doppler anemometry system was employed to determine differences in droplet size distributions, while an optical spray pattern analyzer was used to compare the spray pattern for the various fuel blends as they emerged from the T-56 nozzle.

Influence of a Novel 3D Leading Edge Geometry on the Aerodynamic Performance of Low Pressure Turbine Blade Cascade Vanes

This paper addresses the issue of aerodynamic performance of a novel 3D leading edge modification to a reference low pressure turbine blade. An analysis of tubercles found in nature and used in some engineering applications was employed to synthesize new leading edge geometry. A sinusoidal wave-like geometry characterized by wavelength and amplitude was used to modify the leading edge along the span of a 2D profile, rendering a 3D blade shape. The rationale behind using the sinusoidal leading edge was that they induce streamwise vortices at the leading edge which influence the separation behaviour downstream. Surface pressure and total pressure measurements were made in experiments on a cascade rig. These were complemented with computational fluid dynamics studies where flow visualization was also made from numerical results. The tests were carried out at low Reynolds number of 5.5 × 104 on a well-researched profile representative of conventional low pressure turbine profiles. The performance of the new 3D leading edge geometries was compared against the reference blade revealing a downstream shift in separated flow for the LE tubercle blades; however, total pressure loss reduction was not conclusively substantiated for the blade with leading edge tubercles when compared with the performance of the baseline blade. Factors contributing to the total pressure loss are discussed.Copyright

On the Prediction of Pollutant Emission Indices From Gas Turbine Combustion Chambers

This paper surveys existing emissions models used in the prediction of NOx. The prediction of jet engine emission indices from fundamental principles have proven to be difficult due to the complex physical and chemical interactions occurring within their combustion chambers. Present day prediction of engine emission indices during engine development relies on published models, which are based on limited sets of data measured on older combustion chambers where minimizing pollutant emissions was not a major design criteria. Such empirical and semi-empirical models can, however, provide upper emission limits for new engine designs. A database comprising a wider range of experimental data (over 2000 measured points) taken from the literature was used to test the models. Advanced techniques were applied to optimize the coefficients of proportionality of governing equations of the best models in the literature. Most models tend to consistently over or under predict the measured values. In most cases, even though the standard deviation of the predicted values was not reduced, the correlation error was improved by removing this bias.© 2012 ASME

On the Prediction of Nitrogen Oxides From Gas Turbine Combustion Chambers Using Neural Networks

This paper describes a method of predicting the oxides of nitrogen emissions from gas turbine combustion chambers using neural networks. A short review of existing empirical models is undertaken and the reasoning behind the choice of correlation variables and mathematical formulations is presented. This review showed that the mathematical functions obtained from the underlying theory used to develop the semi-empirical model ultimately limit their general applicability. Under these conditions, obtaining a semi-empirical model with a large domain and good accuracy is difficult. An overview of the use of neural networks as a modelling tool is given. Using over 2000 data points, a neural network that can predict NOx emissions with greater accuracy than published correlations was developed. The coefficients of determination of the prediction for the previous published semi-empirical models are 0.8048 and 0.7885. However one tends to grossly overpredict and the other underpredict. The coefficient of determination is 0.8697 for the model using a neural network. Because of the nature of neural networks, this more accurate model does not allow better insight into the physical and chemical phenomena. It is however, a useful tool for the initial design of combustion chambers.Copyright

Cycle-by-Cycle Exhaust Temperature Monitoring for Detection of Misfiring and Combustion Instability in Reciprocating Engines

This paper describes a means of achieving cycle-by-cycle combustion monitoring of reciprocating engines without the use of cylinder pressure sensors. This approach is intended primarily for engines that are not equipped with indicator passages (that would facilitate the installation of cylinder pressure sensors) but are (or can be) equipped with fittings for individual cylinder exhaust thermocouples. The monitoring system uses rugged exhaust temperature probes and advanced signal processing and analysis to detect cycle-by-cycle variations in exhaust temperatures and correlates these with conventional combustion analysis parameters. The system is particularly useful for detecting the deteriorations in combustion stability that precede misfiring as well as individual misfire events if they occur. Engine test results are presented showing the correlation between the exhaust temperature signal and parameters based upon cylinder pressure measurements. The ability to detect low level combustion instability and isolated, individual misfires has been demonstrated on a 95 liter V12 industrial natural gas engine. It as also been shown that successful acquisition of high fidelity exhaust temperature signals for the combustion analysis can be achieved in the presence of the high levels of electromagnetic interference typical of a power generation facility.Copyright

Experimental Evaluation of Service-Exposed Nozzle Guide Vane Damage in a Rolls Royce A-250 Gas Turbine

A unique methodology and test rig was designed to evaluate the degradation of damaged nozzle guide vanes (NGVs) in a transonic annular cascade in the short duration facility at the Royal Military College. A custom test section was designed which featured a novel rotating instrumentation suite. This permitted 360 deg multispan traverse measurements downstream from unmodified turbine NGV rings from a Rolls-Royce/Allison A-250 turbo-shaft engine. The downstream total pressure was measured at four spanwise locations on both an undamaged reference and a damaged test article. Three performance metrics were developed in an effort to determine characteristic signatures for common operational damage such as trailing edge bends or cracked trailing edges. The highest average losses were observed in the root area, while the lowest occurred closer to the NGV tips. The results from this study indicated that multiple spanwise traverses were required to detect localized trailing edge damage. Recommendations are made for future testing and to further develop performance metrics.

Optical Patternation of Gas Turbine Fuel Sprays During Simulated Engine Operating Conditions

Fuel atomizer condition can have a significant impact on gas turbine hot section component life. In order to investigate the depth of this influence, an experimental test apparatus was constructed, which allowed for optical access to the primary zone of a Rolls-Royce/Allison T56–A–15 turboprop combustion chamber. Test conditions were matched to simulate altitude cruise conditions of a C–130H Hercules military transport aircraft. T56 fuel nozzles of various conditions were tested in free air and then in the test rig using optical patternation techniques. Results indicated that spray characteristics observed in quiescent ambient air persisted under the representative engine operating conditions both burning and non-burning. The optical patternation tests also revealed the influence of combustion liner airflow patterns on the spray within the region of the primary zone that was observed. Conclusions were drawn such as the persistence of spray features observed in open air testing when nozzles were tested at engine representative conditions and recommendations were made for future experimentation.Copyright

Influence of Freestream Turbulence on the Aerodynamic Performance of Transonic Vanes

The present authors have reported a noticeable reduction in the aerodynamic performance of turbine vanes which had been modified as a result of commonly applied repair processes. These tests were done at a low turbulence level to isolate the profile-only effect. In the present research, the effect of the same profile modification on the performance of the vanes was investigated at engine representative flow conditions by increasing the turbulence level and length scale. Since the tested vane profiles in the present research were synthesized using the profile of LPT vanes, the turbulence level was maintained at around 4% and the length scale was set at 2 cm. In the present investigation, calculations with computational fluid dynamics and measurements in a transonic cascade rig were carried out. The high turbulence level in the cascade rig was produced using a passive turbulence-generating grid and in CFD by specifying the desired level and length scale. Coordinates of the baseline profile were obtained from the LPT vanes of an in-service turboshaft engine using 3D optical scanning and digital modeling. The repaired vanes were synthesized using profiles representative of two specific repair types. In both methods, flow visualization was carried out using axial density gradient or schlieren and exit total pressure was obtained numerically or using a multihole probe. Further insight into the flow phenomenon was obtained by surface flow visualization in the cascade rig using a graphite and paraffin oil mixture and by computed surface pressure distributions on the vane. The shock pattern in the cascade for low and high turbulence flows was similar; however, the surface flow pattern exhibited a significant difference for the two conditions. The total pressure ratio and cascade loss also showed some differences.Copyright

Influence of a Novel Three-Dimensional Leading Edge Geometry on the Aerodynamic Performance of Transonic Cascade Vanes

This paper addresses the issue of aerodynamic performance of a novel 3D leading edge modification to a reference vane. An analysis of tubercles found in nature and some engineering applications was used to synthesize new leading edge geometry. Three variations of the reference low pressure turbine vane were obtained by changing the characteristic parameters of the tubercles. Shock structure, surface flow visualization and total pressure measurements were made through experiments in a cascade rig, as well as through computational fluid dynamics. The tests were carried out at design zero incidence and off-design ±10-deg and ±5-deg incidences. The performance of the new 3D leading edge geometries was compared against the reference vane. Some leading edge tubercle configurations were effective at decreasing total pressure losses at positive inlet incidence angles. Numerical results supplemented experimental results.Copyright

Development of a Spark Discharge PM Sensor for Measurement of Engine-Out Soot Emissions

Filter paper methods are well recognized as an effective means of measuring soot emissions from diesel engines. However, these methods provide an average soot value over a relatively long time period, rather than a real-time signal. Real-time measurements of engine-out soot emissions that could track changes in soot levels during transient operating conditions would be useful for the optimization of engine control strategies such as exhaust gas recirculation. This paper presents experimental results obtained using a real-time PM sensor based upon a spark discharge measuring principle. Like traditional filter paper devices, it is sensitive to the carbon or soot component of the particulate matter emitted by diesel engines. The sensor was tested on a turbocharged diesel engine, and compared with reference measurements of Filter Smoke Number (FSN) from an AVL 415s smokemeter. Improvements to the sensor made it possible to measure soot levels at FSN values over 3.5, while retaining good sensitivity below FSN values of 0.1. The sensor signal showed a high correlation with the reference FSN measurements. This correlation was used to develop a signal processing technique so the sensor provided a real-time signal for predicted FSN. Conversion of the FSN values to mass concentration values (using published techniques for the reference instrument) indicated that the output of the spark discharge soot sensor was nearly linear with mass concentration over a substantial portion of the measuring range. The sensor showed a response time of under 2 seconds to step changes in FSN levels.© 2010 ASME