Hermann Haselbacher

Numerical Investigation of the Effect of Ash Particle Deposition on the Flow Field Through Turbine Cascades

Ash deposition on turbine blade surfaces is studied in this work using a particle deposition model. The model involves the three main processes: particle transport to the blade surface particle sticking at the surface and particle detachment from the surface. The model is used to investigate the effect of ash particle deposition on the flow field through turbine cascades. The surface velocity and the downstream total pressure coefficient are calculated for the clean and the fouled blade profiles and used in this investigation. The profile of the clean blade is chosen from the literature for which flow field measurements are available. The two dimensional compressible flow field is solved for the clean blade using the RNG k-e turbulence model with the two layer zonal model for the near-wall region. The results are compared to the experimental data. The flow field is solved at the conditions expected in modern gas turbines. The deposition distribution on the blade surface is calculated during three periods of 12 operating hours each assuming inlet particle concentration as 100 ppmw. The fouled blade profile is predicted after each period. Then the flow field and deposition calculations are repeated to account for the time-dependent particle deposition. The flow field is calculated for the fouled blade after operating hours and investigated using the experimental data and the numerical calculations of the clean blade. The profile loss of the fouled blade is also predicted and compared to that of the clean blade.Copyright

Augmentation of Gas Turbine Power Output by Steam Injection

Steam injection into the combustion chambers of gas turbines (GT) increases their power output. Additionally, the thermal efficiency can be raised, if steam is generated by exhaust heat.The types of steam injected gas turbines (STIG) are distinguished according to the kind of limit to the amount of steam that can be injected. A gas turbine is called partial STIG, if it cannot utilize the total amount of steam that could be generated by the gas turbine exhaust heat. The limit is given by the flow capacity of the turbine. If, on the other hand, the gas turbine is sized such that the entire amount of steam producible can be utilized, it is called full STIG.Three different partial STIG cooling models were selected to analyze the power output, the efficiency and the impact on two important components. Since the differences in the results for the three cycles are marginal, the following conclusion can be briefly summarized: Compressor surge turned out to be the strongest limit for overloading the gas turbine. At the point of maximum overload — where safe operation is still guaranteed — the steam mass flow amounts to one tenth of the nominal compressor air mass flow. At this operating point, the power output can be raised by more than 30% with a simultaneous increase in efficiency.Based on the gas turbine configurations used for the partial STIGs, the preliminary designs of two full STIG cycles have been developed. However, for full STIG operation by injection of the total amount of steam producible, either the compressor or the turbines of the original gas turbine have to be modified. In this case, the steam flow exceeding that required for cooling has to be injected into the compressed air in front of the combustor. Depending on whether the compressor is scaled down or the turbines are scaled up, the power output of full STIGs is 30 to 135% higher than that of the original gas turbine. The gross thermal efficiency is about 50.5.%.Copyright

Effect of Turbulence Modeling on Particle Dispersion and Deposition on Compressor and Turbine Blade Surfaces

One of the main mechanisms that control particle movement is the turbulent diffusion by which the particles in the turbulent boundary layer migrate to the surface under the influence of random flow fluctuations. Theoretical approaches to particle dispersion use random walk models to represent the effect of turbulent fluctuation velocity on particle movement. As a consequence, the turbulence model has a significant effect on the particle trajectory. Particle sticking probability, on the other hand depends upon the particle impact velocity. Moreover, the wall shear stress that is calculated from the turbulence model is the main cause of particle detachment from the surface.In this work, the effect of turbulence models on particle dispersion, deposition on turbine blade surfaces and detachment from the surfaces is studied. Two turbulence models have been tested: the Renormalization Group (RNG) k-e model and the standard k-e model. The near-wall region is solved by two different models: the standard wall function and the two-layer zonal model.It is found that the RNG k-e model with the two-layer zonal near-wall model is the more appropriate turbulence model for particle deposition. It is also concluded that the standard wall function should not be used when solving the flow field near the wall for particle deposition. The reason is that this method does not give the detailed solution of the flow near the wall that is very important for deposition models.Copyright

On the Modeling of Tip Leakage Flow in Axial Turbine Blade Rows

The starting point of this paper is an established turbine tip leakage loss model based on energy considerations. The model requires a discharge coefficient as an empirical input. The discharge coefficient is the ratio of the actual to the theoretical tip gap mass flow rate, The nondimensional parameters influencing the discharge coefficient are determined by a dimensional analysis. These parameters are: gap width to length ratio, end wall speed to gap flow velocity ratio and gap Reynolds number. Ranges for these parameters, valid for typical turbine tip gap situations, are presented. The numerical investigation of the turbulent flow in a plane perpendicular to the blade chord line supplies the discharge coefficient versus the nondimensional gap width. Depending on the gap width to length ratio, various degrees of mixing of the flow downstream of the vena contracta can be detected. Based on these observations, a simple tip gap flow model is presented. The discharge coefficients computed by this model are compared with the numerical results as well as with experimental values from the literature. Finally, the model is used to calculate the discharge coefficients of improved tip gap geometries (squealers, winglets).Copyright

CFD: Analysis of the Influence of Modified Tangential Air Inlets on the Combustion Behavior of a Two-Stage Combustion Chamber

At the Institute of Thermal Turbomachines and Powerplants at the Vienna University of Technology, a two-stage combustion chamber was designed and constructed to directly drive a gas turbine by combustion of wood dust. A commercial CFD-solver was applied to examine the effects of modifications of the geometry on the combustion performance. Since this parameter study was done with the same operation parameters and the same boundary conditions, the computational results represent the influence of the different flow fields caused by the modified combustion chamber. The particle gasification time, the temperature, and the chemical composition of the flue gas at the combustion chamber exit have been used for the assessment of the combustion performance.Copyright

Combustion Chamber for a Directly Wood Particle Fired Gas Turbine

Most of the processes using wood fuels in gas turbine applications that are presently being studied are based on gasification of the wood fuel and operating the gas turbine with the product gas. An alternative is running the gas turbine with the hot gas from a wood combustor — the directly wood particle fired gas turbine. This technique offers the possibility to realise efficient and cost effective small scale power generation systems in the low power range (1–2 MWe). For realizing a directly wood particle fired gas turbine, the Institute of Thermal Turbomachines and Powerplants at the Vienna University of Technology developed a two stage combustor. Solid and liquid fuels require relatively long residence times and good mixing with the oxidant to be completely burned. This can be achieved in the primary stage designed as a cyclone combustor/gasifier. In the cyclone chamber, burning fuel particles are suspended, according to their size, caused by centrifugal and drag forces. This cyclone effect of the flow offers the possibility that big particles remain in the cyclone combustor until they have been completely burned. Using a two stage combustor, the combustion process can be divided into two zones: A primary zone for fuel-rich pyrolysed-gasified-combustion and a secondary zone where the gasification products from the primary zone are oxidized with excess air. Staged combustion has the potential to reduce NOx (NO, NO2 and N2 O), CO and total hydrocarbons Cn Hm concentrations in the exhaust. A large series of test runs was carried out with 3 different fuels, numerous fuel feed rates and equivalence ratios in the cyclone combustor resulting in stable operating conditions and almost total carbon burn-out. The main purpose of the test runs was to investigate the effect of air staging and temperature on the emissions of CO, Cn Hm and NOx .Copyright

CFD-Analysis of the Combustion in the Primary Stage of a Two-Stage Combustion Chamber

Combustion of wood powder may be applied in a two-stage multi-inlet combustion chamber. The primary stage of the combustion chamber has tangential air inlets to provide high swirl flow. The wood powder and its conveying air enter the gasification chamber axially through a center inlet in the bottom. The aim of the investigation was the analysis of the combustion flow of the primary stage of the combustion chamber. The calculation grid was three-dimensional and unstructured. Turbulence was modelled with the Reynolds-Stress-Model, species with mixture fraction/pdf-approach, radiation with the P1-model. Postprocessing has been done for particle tracks, the temperature distribution and tangential velocity distribution and for the species distributions of solid carbon, carbon monoxide, carbon dioxide and oxygen as well.Copyright