Loughborough University
Leicestershire, UK
LE11 3TU
+44 (0)1509 263171
Loughborough University

Loughborough University Institutional Repository

Please use this identifier to cite or link to this item: https://dspace.lboro.ac.uk/2134/13539

Title: Aerodynamic design of the coolant cdelivery system for an intercooled aero gas turbine engine
Authors: A'Barrow, Chris
Keywords: Intercooled
Gas turbine
Ducting
Coolant
Diffusion
OGV
FOGV
Delivery
Aerodynamic
Issue Date: 2013
Publisher: © Chris A'Barrow
Abstract: The Advisory Council of Aeronautical Research in Europe (ACARE) has set record emission reduction targets for 2020, in response to increased awareness of global warming issues and the forecast high level of growth in global air traffic. In order to meet this legislation engine designers have to consider new and unconventional designs. An intercooled aero-engine with a heat exchanger (HX) positioned between the IP and HP compressors has the potential to reduce emissions and/or reduce specific fuel consumption relative to conventional engine cycles. In such an engine a coolant delivery system is required to bleed a proportion of the bypass flow, from behind the fan outlet guide vane (FOGV), rapidly diffuse the flow (to reduce pressure loss through the HX modules) and present it to the intercooler (i.e. heat exchanger) modules for cooling. This spent cooling air is then fed back into the bypass duct. To realise the benefits of the intercooled cycle the coolant delivery system must diffuse the flow, within the geometrical constraints, with minimal pressure loss and present it to the heat exchanger modules with suitable flow characteristics over a range of operating conditions. Therefore, a predominately experimental study, complemented with CFD predictions, was undertaken to investigate the design and performance of a coolant delivery system aimed at providing high pressure recovery in a relatively short length. For this to be achieved some pre-diffusion of the flow is required upstream of the offtake (i.e. by making the offtake larger than the captured streamtube), with a controlled diffuser or hybrid diffuser arrangement located downstream of the offtake. Although targeted at an intercooled aero-engine the concept of a system that produces a high pressure recovery in a limited length is applicable to a variety of applications. Experimental data were obtained on a modified existing low speed isothermal annular test facility operating at nominally atmospheric conditions. The offtake must operate aft of the FOGV in a highly complex flow field environment. Hence, a 1½ stage axial flow compressor (IGV, rotor and modified OGV) was used to simulate the unsteady blade wakes, secondary flows, loss cores and other turbo-machinery features that can significantly influence offtake performance. Preliminary numerical (CFD) studies enabled an offtake configuration to be determined and provided understanding of the governing fluid mechanic processes. A relatively small scale, low speed test facility was designed that had the capability to evaluate aerodynamic processes in isolation (i.e. pre-diffusion, controlled diffusion, hybrid diffusion) and full system modelling to enable the complex interaction between these flow processes to be assessed. Hence an optimal system could be characterised in terms of total pressure loss, static pressure recovery and flow profiles at HX inlet. Measurements and numerical predictions are initially presented for a baseline configuration with no offtake present. This enabled the OGV near field region to be characterised and provided a datum, relative to which the effects of introducing an offtake could be assessed. The results showed that in the near field region (i.e. within one chord downstream of the FOGV) the high velocity gradients in the circumferential direction, and associated turbulent shear stresses, dominate the profile mixing and loss production. There is little mixing out of profiles in the radial direction. Furthermore, the relatively large amount of kinetic energy associated with the compressor efflux and its subsequent mixing to a more uniform profile (i.e. reduced blockage) results in a significant static pressure recovery (Cp=5.5%). With the offtake present a variety of configurations were investigated including different levels of pre-diffusion, prior to the offtake, and different offtake positions. This enabled evaluation of the upstream pressure effects and interaction with the upstream FOGV. For very compact systems of short length, such that the gap between the OGV and offtake is relatively small, the amount of pre-diffusion achievable is limited by the offtake pressure field and its impact on the upstream OGV row. This pressure field is also influenced by parameters such as the non-dimensional offtake height and splitter thickness. For systems of increased length a significant amount of flow pre-diffusion can be achieved with little performance penalty (relative to the datum configuration). Hence, the loss associated with mixing blade wakes and secondary flows in an adverse pressure gradient is relatively small. However, the pre-diffusion level is eventually limited, to approximately 1.5, by the increased distortion and pressure losses associated with the captured streamtube. Further measurements were made with various controlled diffuser and hybrid diffusers (of varying area ratio) downstream of the offtake and various levels of pre-diffusion. The flow profile that is presented to the controlled diffuser is directly influenced by the upstream pre-diffusion process. Hence, in this case the upstream-downstream interaction is relatively strong. Conversely, the downstream-upstream interaction, between the controlled diffuser and pre-diffusion process, is relatively weak and thus has little effect on the upstream flow field. The data enabled an optimal system to be characterised (pre-diffusion/controlled diffusion split) in terms of total pressure loss, static pressure recovery and flow profiles at HX inlet. A total system diffusion of 1.8 was achievable with a pre-diffusion of 1.4 and controlled diffusion of 1.25, with further increases in either the pre-diffusion level or the controlled diffuser area ratio destabilising the system. This was achieved with an absolute mass weighted total pressure loss of 11% measured from FOGV inlet to the controlled diffuser exit plane. Utilising a hybrid bled diffuser, combined with the pre-diffusion, enabled a total system diffusion of 2.24 to be achieved. The system incorporated a 6% bleed from the hybrid diffuser and a system total pressure loss of 13%. Experimental and computational results obtained in the current research have provided an understanding of the governing flow mechanisms and quantified the geometric and aerodynamic interaction of the offtake with the FOGV and between the diffusion processes. This has enabled a design methodology to be outlined that provides approximate information on system geometry and performance (in terms of optimal diffusion split and total pressure loss) for future coolant delivery systems with minimal effort. Preliminary design maps have been developed to define the magnitude of the interaction between the offtake and FOGV in terms of the offtake height, pre-diffusion level, the splitter thickness and the axial distance between the fan OGV and offtake. In this way systems of optimal diffusion split, minimum pressure loss and minimal axial length can be determined.
The Advisory Council of Aeronautical Research in Europe (ACARE) has set record emission reduction targets for 2020, in response to increased awareness of global warming issues and the forecast high level of growth in global air traffic. In order to meet this legislation engine designers have to consider new and unconventional designs. An intercooled aero-engine with a heat exchanger (HX) positioned between the IP and HP compressors has the potential to reduce emissions and/or reduce specific fuel consumption relative to conventional engine cycles. In such an engine a coolant delivery system is required to bleed a proportion of the bypass flow, from behind the fan outlet guide vane (FOGV), rapidly diffuse the flow (to reduce pressure loss through the HX modules) and present it to the intercooler (i.e. heat exchanger) modules for cooling. This spent cooling air is then fed back into the bypass duct. To realise the benefits of the intercooled cycle the coolant delivery system must diffuse the flow, within the geometrical constraints, with minimal pressure loss and present it to the heat exchanger modules with suitable flow characteristics over a range of operating conditions. Therefore, a predominately experimental study, complemented with CFD predictions, was undertaken to investigate the design and performance of a coolant delivery system aimed at providing high pressure recovery in a relatively short length. For this to be achieved some pre-diffusion of the flow is required upstream of the offtake (i.e. by making the offtake larger than the captured streamtube), with a controlled diffuser or hybrid diffuser arrangement located downstream of the offtake. Although targeted at an intercooled aero-engine the concept of a system that produces a high pressure recovery in a limited length is applicable to a variety of applications. Experimental data were obtained on a modified existing low speed isothermal annular test facility operating at nominally atmospheric conditions. The offtake must operate aft of the FOGV in a highly complex flow field environment. Hence, a 1½ stage axial flow compressor (IGV, rotor and modified OGV) was used to simulate the unsteady blade wakes, secondary flows, loss cores and other turbo-machinery features that can significantly influence offtake performance. Preliminary numerical (CFD) studies enabled an offtake configuration to be determined and provided understanding of the governing fluid mechanic processes. A relatively small scale, low speed test facility was designed that had the capability to evaluate aerodynamic processes in isolation (i.e. pre-diffusion, controlled diffusion, hybrid diffusion) and full system modelling to enable the complex interaction between these flow processes to be assessed. Hence an optimal system could be characterised in terms of total pressure loss, static pressure recovery and flow profiles at HX inlet. Measurements and numerical predictions are initially presented for a baseline configuration with no offtake present. This enabled the OGV near field region to be characterised and provided a datum, relative to which the effects of introducing an offtake could be assessed. The results showed that in the near field region (i.e. within one chord downstream of the FOGV) the high velocity gradients in the circumferential direction, and associated turbulent shear stresses, dominate the profile mixing and loss production. There is little mixing out of profiles in the radial direction. Furthermore, the relatively large amount of kinetic energy associated with the compressor efflux and its subsequent mixing to a more uniform profile (i.e. reduced blockage) results in a significant static pressure recovery (Cp=5.5%). With the offtake present a variety of configurations were investigated including different levels of pre-diffusion, prior to the offtake, and different offtake positions. This enabled evaluation of the upstream pressure effects and interaction with the upstream FOGV. For very compact systems of short length, such that the gap between the OGV and offtake is relatively small, the amount of pre-diffusion achievable is limited by the offtake pressure field and its impact on the upstream OGV row. This pressure field is also influenced by parameters such as the nondimensional offtake height and splitter thickness. For systems of increased length a significant amount of flow pre-diffusion can be achieved with little performance penalty (relative to the datum configuration). Hence, the loss associated with mixing blade wakes and secondary flows in an adverse pressure gradient is relatively small. However, the pre-diffusion level is eventually limited, to approximately 1.5, by the increased distortion and pressure losses associated with the captured streamtube. Further measurements were made with various controlled diffuser and hybrid diffusers (of varying area ratio) downstream of the offtake and various levels of prediffusion. The flow profile that is presented to the controlled diffuser is directly influenced by the upstream pre-diffusion process. Hence, in this case the upstreamdownstream interaction is relatively strong. Conversely, the downstream-upstream interaction, between the controlled diffuser and pre-diffusion process, is relatively weak and thus has little effect on the upstream flow field. The data enabled an optimal system to be characterised (pre-diffusion/controlled diffusion split) in terms of total pressure loss, static pressure recovery and flow profiles at HX inlet. A total system diffusion of 1.8 was achievable with a pre-diffusion of 1.4 and controlled diffusion of 1.25, with further increases in either the pre-diffusion level or the controlled diffuser area ratio destabilising the system. This was achieved with an absolute mass weighted total pressure loss of 11% measured from FOGV inlet to the controlled diffuser exit plane. Utilising a hybrid bled diffuser, combined with the prediffusion, enabled a total system diffusion of 2.24 to be achieved. The system incorporated a 6% bleed from the hybrid diffuser and a system total pressure loss of 13%. Experimental and computational results obtained in the current research have provided an understanding of the governing flow mechanisms and quantified the geometric and aerodynamic interaction of the offtake with the FOGV and between the diffusion processes. This has enabled a design methodology to be outlined that provides approximate information on system geometry and performance (in terms of optimal diffusion split and total pressure loss) for future coolant delivery systems with minimal effort. Preliminary design maps have been developed to define the magnitude of the interaction between the offtake and FOGV in terms of the offtake height, pre-diffusion level, the splitter thickness and the axial distance between the fan OGV and offtake. In this way systems of optimal diffusion split, minimum pressure loss and minimal axial length can be determined.
Description: A Doctoral Thesis. Submitted in partial fulfilment of the requirements for the award of Doctor of Philosophy of Loughborough University.
Sponsor: Rolls-Royce
URI: https://dspace.lboro.ac.uk/2134/13539
Appears in Collections:PhD Theses (Aeronautical and Automotive Engineering)

Files associated with this item:

File Description SizeFormat
Form-2013-ABarrow.pdf146.75 kBAdobe PDFView/Open
Thesis-2013-ABarrow.pdf13.43 MBAdobe PDFView/Open

 

SFX Query

Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.