No abstract

To decrease NOx emissions from combustion systems, lean premixed combustion is used. A disadvantage is the higher sensitivity to combustion instabilities, leading to increased sound pressure levels in the combustor and resulting in an increased excitation of the surrounding structure: the liner. This causes fatigue, which limits the lifetime of the combustor. This paper presents a joint experimental and numerical investigation of this acoustoelastic interaction problem for frequencies up to 1kHz. To study this problem experimentally, a test setup has been built consisting of a single burner, 500kW, 5bar combustion system. The thin structure (liner) is contained in a thick pressure vessel with optical access for a traversing laser vibrometer system to measure the vibration levels of the liner. The acoustic excitation of the liner is measured using pressure sensors measuring the acoustic pressures inside the combustion chamber. For the numerical model, the finite element method with full coupling between structural vibration and acoustics is used. The flame is modeled as an acoustic volume source corresponding to a heat release rate that is frequency independent. The temperature distribution is taken from a Reynolds averaged Navier Stokes (RaNS) computational fluid dynamics (CFD) simulation. Results show very good agreement between predicted and measured acoustic pressure levels. The predicted and measured vibration levels also match fairly well.

No abstract

No abstract

A numerical case study on a HPT-shroud of a medium-sized commercial engine has been carried out to investigate the heat loading and the possible redistribution (number of channels, position and exit angle) of shroud cooling channels facing the turbine blade tip. A combination of modeling vehicles was used to quantify the aerodynamics, the thermodynamics and resulting heat loads on the shroud. This includes a 1-D gas turbine performance simulation model, engineering models for cooling flow distributions and heat loads, CFD modeling of the HPT flow including some tip flow effects and the finite element modeling to calculate the temperature and stress distribution in the solid shroud. Regions with high temperatures and/or maximum thermal stresses and the potential for reduction by relocating the cooling channels at equal amounts of cooling flow were identified. Although the physics involved in the processes is much more complicated than modeled, the parametric studies gave valuable insight and quantitative results in terms of differences in shroud temperatures and thermal stresses. A complementary experimental study on shroud maintenance and service experiences (not published yet) has delivered data for model input support and comparison with the numerical results.

The subject of the Desire project is the combustion chamber, as depicted in figure 1.2. One of the combustion chamber designs currently used in gas turbines is the so-called annular combustor (figure 1.3). In this design the combustion chamber is shaped as an annulus in which multiple burners (for instance 24) cause multiple swirling flames. It is very expensive to build a full annulus as a test rig, and therefore one section, containing one burner, is taken from the annulus. The interaction between flames from different burners is Combustion, acoustics and vibrationThis section introduces the basic phenomenology, which is schematically depicted in figure 1.4, starting with the flame. The most characteristic influence of the flame is that it generates a temperature field (1), which strongly influences the acoustics of the system. This is mostly a static influence. The steady state temperature field created by the flame determines the steady state temperature field for the acoustics of the system. Unsteady combustion can give time-dependent temperature differences. When the flame moves, the local temperature field in the flame zone changes, but these perturbations have no global influence. When the amount of fuel that comes to the flame changes, the temperature of the total flame changes. These perturbations propagate through the combustion chamber with the flow velocity and are called entropy waves [78]. These are not taken into account in this thesis. Besides the steady influence, the flame also has unsteady components. Because the combustion is turbulent, the combustion speed constantly changes around the mean value. These changes generate an acoustic volume source (2), which gives an acoustic field in the combustor (this is commonly called combustion roar ). This acoustic field in turn generates perturbations in the fuel and air flow to the burner and to the mixture coming to the flame (3), which again generates perturbations in the speed of combustion. This loop can become unstable, which is known as a combustion instability. This behavior is studied extensively in the literature [23,62,69]. Another well-known example of an unstable feedback loop between heat and acoustics is the so-called Rijke tube [82].Vibration of the structure is influenced directly by the flame through the temperature field (1). This is again a steady phenomenon, the steady flame determines the temperature of the liner and therefore its material properties. Changes in operating conditions cause changes in thermal expansion of components and therefore cyclic thermal stresses, which can directly lead to relatively low cycle fatigue [95]. The interaction between acoustics and vibration is similar to that between combustion and acoustics. The acoustic field acts as a pressure load on the structure (4), which vibrates in response. This vibration imposes a velocity boundary condition on the acoustic domain (5), which generates an acoustic field. This loop does not become unstable, because there Table A.1: Coefficients of the c p polynomial, g...

No abstract