The initial design of a fluidically controlled variable geometry fuel injector for gas turbine combustion systems

Abstract

Listen

Translate article

The initial design of a fluidically controlled variable geometry fuel injector for use within a gas turbine combustion system is reported. Combustion fundamentals are examined to establish the requirements of a variable geometry combustion system. Variable geometry fuel injectors potentially offered the most control over the combustion process in order to improve stability at low power whilst maintaining low emissions at higher power conditions. Mechanical variable geometry fuel injectors were tested to establish the potential of the technique. Improved stability of around 10% was shown with a small modulation in the airflow. Fluidics offers a no moving parts solution to variable geometry. Fluidic devices have been utilised in engineering for several decades, and many devices have been conceived. The design of the fuel injector was based on an existing fluidic device developed for oil separation techniques. Testing of the device showed that it could meet the requirements of a variable geometry fuel injector, however the pressure loss was too high for direct application into a gas turbine. Pressure loss requirements were paramount in the overall design, in order to maximise specific fuel consumption and avoid compressor stall/surge. The design was modified with the use of computational fluid dynamics, in order to reduce the pressure loss to 4.5%. British Crown Copyright © 1998/DERA Published by the American Institute of Aeronautics and Astronautics, Inc., with permissions. | Introduction Current gas turbine engine trends are towards increased thrust/weight ratio which requires the engine to perform at higher operating compression ratios and thus higher combustor inlet temperatures. The combustor will also be expected to run richer at the high power end of the cycle, which combined with the increased inlet temperature, will increase flame temperatures. NOx emissions produced at high power conditions are mainly due to thermal NOx mechanisms [1], and are proportional to flame temperature and residence time. Future cycles with higher inlet temperatures and richer AFR (at high power) will inevitably have higher emissions of both NOx and smoke. This will impact on civil gas turbine combustors due to expected NOx legislation and certification requirements, but will also affect the stealth properties of military aircraft (smoke and nitrogen dioxide (NO2) being visible). To minimise NOx and smoke emissions at full power, the primary zone (shown in Figure 1) must be as weak as possible. However, this makes stability poor at low power conditions, and altitude relight becomes hard to achieve. Thus a move away from traditional designs is required. Variable Geometry Requirements The combustion is initiated and stabilises in the primary zone (Figure 1). High power conditions requires good mixing and atomisation, and lean localised AFRs in order to ensure smoke and other emissions levels are low and that the combustor exit traverse is acceptable. However, low power stability benefits from rich areas within the primary zone of the combustor, which enables combustion to occur when the overall air/fuel ratio is much weaker than the flammability limit of kerosene. Figure 1 Generic Gas Turbine Combustor In traditional combustion systems rich regions can occur in the combustor due to poor mixing when the fuel is not mixed into the air-stream fully, and through poor atomisation resulting in large droplets being formed. Another requirement for good low power stability, is low velocity airflow and high recirculation. This is essential to allow the low intensity flame to burn without being blown out of the combustor by high air velocities, and to allow hot combustion gases to recirculate enhancing fuel vaporisation. The ability of the combustion system to sustain a flame in the vicinity of high velocity air is dependent on many variables. One of the most important variable is the laminar burning velocity (uj). The burning velocity of a kerosene spray is dependent on the local concentration of fuel in both the liquid and vapour phase. Considering a mixture where the fuel is fully vaporised and mixed with the air, the burning velocity exhibits a characteristic typical of that shown in Figure 2 [2], which shows equivalence ratio (actual Fuel Air Ratio (FAR)/Stoichiometric FAR) against burning velocity. As more fuel is added to the weak mixture, the burning velocity increases, reaches a peak at slightly richer than stoichiometric and then reduces until the rich limit is reached (approximately 3:1 APR for kerosene).