Evaluation of a 3D Viscous Code for Turbomachinery Flows

Author(s):  
W. John Calvert ◽  
Andrew W. Stapleton ◽  
Paul R. Emmerson ◽  
Cecil R. Buchanan ◽  
Christopher M. Nott

Computational Fluid Dynamics (CFD) codes play an increasingly important role in the design and development of turbomachinery for modern gas turbine engines. As a result additional emphasis is being placed on the evaluation of the codes to ensure that they are working correctly and to indicate the accuracy which is likely to be achieved in practice. At DERA a programme of work has been carried out to evaluate the TRANSCode 3D viscous flow code, which was developed from the BTOB3D solver written by Dawes in 1986. A three part strategy for the validation and calibration of the code was adopted, covering comparisons with boundary layer test cases, Q3D compressor cascades and full 3D cases. The results indicated that the grids currently employed for turbomachinery flows limit the accuracy achieved for cases where there is significant laminar flow. For turbulent flows the Baldwin-Lomax turbulence model gives reasonably accurate results for 2D flows in near equilibrium, but it is less satisfactory for more complex flows, when the concept of a simple 2D boundary layer does not apply, and for strong shock wave/boundary layer interactions. Overall it is considered that the code is a valuable tool for turbomachinery analysis and design, but solutions must be assessed with care. Alternative turbulence models and other developments are being pursued for future versions of the code.

2005 ◽  
Author(s):  
Sowjanya Vijiapurapu ◽  
Jie Cui

The Reynolds averaged Navier-Stokes (RANS) equations were solved along with three turbulence models, namely κ-ε, κ-ω, and Reynolds stress models (RSM), to study the fully developed turbulent flows in circular pipes roughened by repeated square ribs. The spacing between the ribs was varied to form three representative types of surface roughness; d–type, intermediate, and k–type. Solutions of these flows at two Reynolds numbers were obtained using the commercial computational fluid dynamics (CFD) software Fluent. The numerical results were validated against experimental measurements and other numerical data published in literature. Extensive investigation of effects of rib spacing and Reynolds number on the pressure and friction resistance, flow and turbulence distribution was presented. The performance of three turbulence models was also compared and discussed.


Author(s):  
M. A. Mawid ◽  
C. A. Arana ◽  
B. Sekar

An advanced thermal management analysis tool, named Advanced Thermal Hydraulic Energy Network Analyzer (ATHENA), has been used to simulate a fuel system for gas turbine engines. The ATHENA tool was modified to account for JP-8/dodecane fuel properties. The JP-8/dodecane fuel thermodynamic properties were obtained from the SUPERTRAP property program. A series of tests of a fuel system simulator located at the Air Force Research Laboratory (AFRL)/Wright Patterson Air Force Base were conducted to characterize the steady state and dynamic behavior of the fuel system. Temperature, pressures and fuel flows for various fuel pump speeds, pressure rise and flow control valve stem positions (orifice areas), heat loads and engine fuel flows were measured. The predicted results were compared to the measured data and found to be in excellent agreement. This demonstrates the capability of the ATHENA tool to reproduce the experimental data and, consequently, its validity as an analysis tool that can be used to carry out analysis and design of fuel systems for advanced gas turbine engines. However, some key components in the fuel system simulator such as control components, which regulate the engine fuel flow based on predetermined parameters such as fan speed, compressor inlet and exit pressures and temperatures, combustor pressure, turbine temperature and power demand, were not simulated in the present investigation due to their complex interactions with other components functions. Efforts are currently underway to simulate the operation of the fuel system components with control as the engine fuel flow and power demands are varied.


Author(s):  
Shaowen Chen ◽  
Qinghe Meng ◽  
Yueqi Liu ◽  
Hongyan Liu ◽  
Songtao Wang

Abstract The most important flow behaviour of supersonic compressor cascades is the shock wave boundary layer interaction (SWBLI). Large eddy simulation (LES) and multiple analysing methods are applied in current study to capture more details of the flow field. It is noted that the LES can catch the dual peaks feature near the SWBLI region with respect to the experimental results. Besides, SWBLI is not only the main losses source in the cascade, but also the most important origin of the unsteadiness behaviour. The high frequency signals correspond to the coherent structure in the boundary layer and dissipate downstream in the cascade, while the low frequency signals relate to the motion of the reflection point of the passage oblique shock wave and dominate the frequency spectrum downstream.


1978 ◽  
Vol 100 (2) ◽  
pp. 232-236 ◽  
Author(s):  
H. E. Weber

A simple, semiempirical method for calculating the laminar, transition, and turbulent boundary layer with arbitrary free stream pressure gradient is developed. Good correlation is obtained with data on general two dimensional turbulent flows, diffuser flows, and the cylinder in cross-flow. However only for the diffuser has the boundary layer flow been coupled with the potential core so that only the inlet conditions and geometry are required. In other cases the free stream velocity distribution must be known or calculable. Skin friction coefficient, momentum thickness Reynolds number, and free stream pressure gradient parameter correlation employs a simple lag theory. With the integral momentum equation the complete boundary layer parameters are obtained as functions of the distance along a surface.


2019 ◽  
Vol 141 (2) ◽  
Author(s):  
Peter R. Forsyth ◽  
David R. H. Gillespie ◽  
Matthew McGilvray

The ingestion and deposition of solid particulates within gas turbine engines has become a very significant concern for both designers and operators in recent times. Frequently aircraft are operated in environments where sand, ash, dust, and salt are present, which can drive damage mechanisms from long term component degradation to in-flight flame-out. Experiments are presented to assess deposition characteristics of sodium chloride (NaCl) at gas turbine secondary air system temperature conditions in horizontal pipe flow. Monodisperse NaCl particles were generated in the size range 2.0–6.5 µm, with gas temperatures 390–480 °C, and metal temperatures 355–730 °C. Two engine-representative surface roughnesses were assessed. An experimental technique for the measurement of deposited NaCl based on solution conductivity was developed and validated. Experiments were carried out under isothermal and nonisothermal/thermophoretic conditions. An initial experimental campaign was conducted under ambient and isothermal conditions; high temperature isothermal results showed good similarity. Under thermophoretic conditions, deposition rates varied by up to several orders of magnitude compared to isothermal rates.


Author(s):  
Sassan Etemad ◽  
Bengt Sunde´n

The developing turbulent flow in a rectangular-sectioned curved duct with an aspect-ratio of 6 was explored using linear and non-linear high- and low-Re k-ε turbulence models. The impact of the inlet flow conditions and grid density was also studied. It was found that the boundary layer thickness upstream the bend has fundamental influence on the secondary flow, velocity profile and turbulence level. The flow in the straight inlet duct was nearly 2-dimensional. The predicted data using Chen’s model and the quadratic high-Re model agreed well with the experimental data from Kim and Patel [1]. Also Suga’s cubic low-Re model performed well. The quadratic low-Re model, however, predicted thicker boundary layer which gave magnified secondary flow with high levels of shear and turbulence. It is not certain that the used high-Re models would perform well in a fully three-dimensional flow. Suga’s model, however, might perform well in other more complex flows.


Author(s):  
Marco Pellegrini ◽  
Hiroshi Endo ◽  
Hisashi Ninokata

A modification of the damping functions in the turbulent dissipation rate transport equations is proposed based on the study given by [1]. The modification accounts for a generalization of the applicability to complex geometries and complex flows. The validation of the proposed model is carried out with Computational Fluid Dynamics (CFD) against a fully developed pipe flow and sharp-bent pipe, both at transitional Reynolds numbers where common turbulence models fail in an accurate predictions of the flow and turbulent characteristics. The results achieved for both geometries demonstrate that the present formulation is able to provide higher agreement and accuracy in the prediction of the flow, in relation to velocity and turbulence characteristics. It has been concluded therefore that the present formulation shows a higher generality in comparison with previously available models, accounting for a more reliable applicability and wider number of cases.


Sign in / Sign up

Export Citation Format

Share Document