Model Setup for the Investigation of Flow Phenomena in a 1.5-Stage Gas Turbine Regarding Hot Gas Ingestion

Abstract

Abstract This paper presents different modeling approaches on rim seal flow in a 1.5-stage test turbine resulting in the analysis of numerical data in comparison to existing test data. The analysis focuses on a very simple axial sealing gap and a reference operation point which shows significant hot gas ingestion in the experiments. The presented model will be used to gather more information on flow phenomena and their effect on the ingestion and hence contributes to a deeper understanding. In combination with broader research this understanding will help manufacturers to reduce their secondary air mass flow rate and increase gas turbine’s efficiency. Experiments on the hot gas ingestion phenomenon have already been carried out at many international test facilities. As part of a recommissioning of a test rig for investigating hot gas ingestion into the wheel side cavities of axial-flow gas turbines at the RWTH Aachen University, this publication presents the numerical model set up in parallel. In particular, the aim is to investigate the sensitivity of the modeled flow field in the wheel side cavity for changes in the modeling and boundary conditions. In addition to the investigation of the sensitivity, the results are compared with existing measurements in order to be able to classify and evaluate the sensitivities in a specific manner. The investigation starts with a mesh study followed by the analysis of different turbulence models including different inlet turbulence intensities. After analyzing different time steps in transient and steady state simulations, calculations are shown concerning different operation points. For all chosen combinations of settings and boundary conditions the main flow is modeled reasonably well compared to experimental data. Also the operation point dependency of the pressure field is modeled well with simple approaches. For low purge air mass flow rates the flow field in the wheel side chamber, consisting of pressure and sealing gas concentration, is only represented with strong deviations in the one-segment model. These deviations are due to large structures forming in the cavity, which can not be reproduced using one or two pitches of the turbine. Above a certain dimensionless purge air mass flow rate, a transition in the flow field can be observed analyzing the experimental data. Above this flow rate the large structures break down and the sealing effectiveness increases. At these operation points the selected numerical model shows good agreement with experimental data regarding pressure and hot gas concentration. To improve predictions at low purge air flow rate multiple segment models will be used in the future.