The Simulation Study of Turbulence Models for Conjugate Heat Transfer Analysis of a High Pressure Air-Cooled Gas Turbine

Abstract

The different turbulence models are adopted to simulate NASA-MarkII high pressure air-cooled gas turbine. The experimental work condition is Run 5411. The paper researches that the effect of different turbulence models for the flow and heat transfer characteristics of turbine. The turbulence models include: the laminar turbulence model, high Reynolds number k-ε turbulence model, low Reynolds number turbulence model (k-ω standard format, k-ω-SST and k-ω-SST-γ-θ) and B-L algebra turbulence model which is adopted by the compiled code. The results show that the different turbulence models can give good flow characteristics results of turbine, but the heat transfer characteristics results are different. Comparing to the experimental results, k-ω-SST-θ-γ turbulence model results are more accurate and can simulate accurately the flow and heat transfer characteristics of turbine with transition flow characteristics. But k-ω-SST-γ-θ turbulence model overestimates the turbulence kinetic energy of blade local region and makes the heat transfer coefficient higher. It causes that local region temperature is higher. The results of B-L algebra turbulence model show that the results of B-L model are accurate besides it has 4% temperature error in the transition region. As to the other turbulence models, the results show that all turbulence models can simulate the temperature distribution on the blade pressure surface except the laminar turbulence model underestimates the heat transfer coefficient of turbulence flow region. On the blade suction surface with transition flow characteristics, high Reynolds number k-ε turbulence model overestimates the heat transfer coefficient and causes the blade surface temperature is high about 90K than the experimental result. Low Reynolds number k-ω standard format and k-ω-SST turbulence models also overestimate the blade surface temperature value. So it can draw a conclusion that the unreasonable choice of turbulence models can cause biggish errors for conjugate heat transfer problem of turbine. The combination of k-ω-SST-γ-θ model and B-L algebra model can get more accurate turbine thermal environment results. In addition, in order to obtain the affect of different turbulence models for gas turbine conjugate heat transfer problem. The different turbulence models are adopted to simulate the different computation mesh domains (First case and Second case). As to each cooling passages, the first case gives the wall heat transfer coefficient of each cooling passages and the second case considers the conjugate heat transfer course between the cooling passages and blade. It can draw a conclusion that the application of heat transfer coefficient on the wall of each cooling passages avoids the accumulative error. So, for the turbine vane geometry models with complex cooling passages or holes, the choice of turbulence models and the analysis of different mesh domains are important. At last, different turbulence characteristic boundary conditions of turbine inner-cooling passages are given and K-ω-SST-γ-θ turbulence model is adopted in order to obtain the effect of turbulence characteristic boundary conditions for the conjugate heat transfer computation results. The results show that the turbulence characteristic boundary conditions of turbine inner-cooling passages have a great effect on the conjugate heat transfer results of high pressure gas turbine.