This paper presents the transient aero-thermal analysis of a gas turbine internal air system through an engine flight cycle featuring multiple fluid cavities that surround a HP turbine disk and the adjacent structures. Strongly1 γ least-squares problem solution ∆ difference δ wall temperature perturbation factor θ time discretization control parameter Superscripts () n n-th time step INTRODUCTIONAccurate prediction of aerodynamic, aero-mechanical and thermo-mechanical phenomena attracts an increasing interest from the gas turbine industry. Efficient and robust analysis procedures able to properly describe the thermal and flow environment within a secondary air system may lead to substantial gains in overall engine performance, weight and components reliability by offering improved means of optimizing designs [1].Typically, in thermal modeling, the internal air system is modeled with user-specified boundary conditions, while more accurate CFD predictions, if available, are applied only on some portions of the system. The user-specified boundary conditions rely heavily on correlations, they require a significant effort from an end-user to correctly represent the complex physical phenomena over a wide range of operating conditions. As the geometries of modern air systems grow in complexity, current models with correlations become painful to build and increasingly obsolete.A general trend in computational modeling of modern turbo-machinery systems is to move towards the "virtual" or "whole" engine simulation [2]. As far as modeling of the internal air systems is concerned, a natural and incremental advance in both the complexity and the accuracy of current modeling is to include and interconnect multiple CFD domains within a single simulation. Among the advantages it may offer are a lower level of human intervention and time required to set up the models, and more importantly, automatic generation of the boundary conditions for the downstream components. However, this comes at a price of a considerably higher computational effort required to run a simulation through an engine transient flight cycle leading to long analysis times.Many studies in recent years sought to improve the predictive capabilities of thermo-mechanical analysis codes by coupling FE solvers to detailed CFD models of individual components to more accurately evaluate wall temperature distribution in turbine cavities, see, for example, [3,4,5,6,7]. While these studies were able to obtain only a general agreement with the experimental data, they did demonstrate many of the fundamental features outlined in earlier investigations. Subsequent efforts attempted to further improve the agreement by including some of both fluid and solid domains 3D geometrical features in the analysis [8] or the effects of solid domain thermo-mechanical distortion on flow dynamics [9]. Still, accurate and automatic predictions of heat transfer in the internal air systems remain a difficult challenge.The main goal of this paper is to provide a snapshot of the state-...

The accuracy of computational fluid dynamics (CFD) for the prediction of flow and heat transfer in a direct transfer pre-swirl system is assessed through a comparison of CFD results with experimental measurements. Axisymmetric and three dimensional (3D) sector CFD models are considered. In the 3D sector models, the pre-swirl nozzles or receiver holes are represented as axisymmetric slots so that steady state solutions can be assumed. A number of commonly used turbulence models are tested in three different CFD codes, which were able to capture all of the significant features of the experiments. Reasonable quantitative agreement with experimental data for static pressure, total pressure and disc heat transfer is found for the different models, but all models gave results which differ from the experimental data in some respect. The more detailed 3D geometry did not significantly improve the comparison with experiment, which suggested deficiencies in the turbulence modelling, particularly in the complex mixing region near the pre-swirl nozzle jets. The predicted heat transfer near the receiver holes was also shown to be sensitive to near-wall turbulence modelling. Overall, the results are encouraging for the careful use of CFD in pre-swirl-system design.

Design of pre-swirl systems is important for the secondary air cooling system of gas turbine engines. In this paper, three pre-swirl nozzles, a cascade vane and two drilled nozzles are analysed and their performances are compared. The two drilled nozzles considered are a straight drilled nozzle and an aerodynamically designed nozzle. CFD analyses are presented for stand-alone and pre-swirl system 3D sector models at engine operating conditions near to engine maximum power condition rotational Reynolds number (Reφ) up to 4.6 ! 107. Nozzle performance is characterised by the nozzle discharge coefficient (CD), nozzle velocity coefficient (η) and cooling air delivery temperature. Two commonly used eddy viscosity models are employed for the study, the standard k-ε and Spalart-Allmaras models with wall functions. Both models give very similar results for CD and η, and are in reasonable agreement with available experimental data. Effects of nozzle or vane number and sealing flow have been analysed. The cascade vanes perform slightly better than the aerodynamically designed drilled nozzles but the final design choice will depend on other component and manufacturing costs. An elementary model is presented to separate temperature losses due to the nozzle, stator drag and sealing flow.

The accuracy of computational fluid dynamics (CFD) for the prediction of flow and heat transfer in a direct transfer preswirl system is assessed through a comparison of CFD results with experimental measurements. Axisymmetric and three-dimensional (3D) sector CFD models are considered. In the 3D sector models, the preswirl nozzles or receiver holes are represented as axisymmetric slots so that steady state solutions can be assumed. A number of commonly used turbulence models are tested in three different CFD codes, which were able to capture all of the significant features of the experiments. A reasonable quantitative agreement with experimental data for static pressure, total pressure, and disk heat transfer is found for the different models, but all models gave results that differ from the experimental data in some respect. The more detailed 3D geometry did not significantly improve the comparison with experiment, which suggests deficiencies in the turbulence modeling, particularly in the complex mixing region near the preswirl nozzle jets. The predicted heat transfer near the receiver holes was also shown to be sensitive to near-wall turbulence modeling. Overall, the results are encouraging for the careful use of CFD in preswirl-system design.

The prediction of the preswirl cooling air delivery and disk metal temperature are important for the cooling system performance and the rotor disk thermal stresses and life assessment. In this paper, standalone 3D steady and unsteady computation fluid dynamics (CFD), and coupled FE-CFD calculations are presented for prediction of these temperatures. CFD results are compared with previous measurements from a direct transfer preswirl test rig. The predicted cooling air temperatures agree well with the measurement, but the nozzle discharge coefficients are under predicted. Results from the coupled FE-CFD analyses are compared directly with thermocouple temperature measurements and with heat transfer coefficients on the rotor disk previously obtained from a rotor disk heat conduction solution. Considering the modeling limitations, the coupled approach predicted the solid metal temperatures well. Heat transfer coefficients on the rotor disk from CFD show some effect of the temperature variations on the heat transfer coefficients. Reasonable agreement is obtained with values deduced from the previous heat conduction solution.