39
IRUS Total
Downloads
  Altmetric

Experiments and modelling of surge in turbocharger compressors

File Description SizeFormat 
Gharaibeh-K-2020-PhD-Thesis.pdfThesis9.79 MBAdobe PDFView/Open
Title: Experiments and modelling of surge in turbocharger compressors
Authors: Gharaibeh, Karim
Item Type: Thesis or dissertation
Abstract: Surge is a mode of flow instability in compressors that has plagued turbomachinery for almost a century. It is a system-level effect excited by the compressor and characterised by oscillations in mass flow and pressure in the through-flow direction. Currently, the lack of confidence in surge modelling accuracy forces engine designers to avoid large areas of the compressor map, potentially inhibiting engine performance. Indeed, industry’s need to better understand and model surge in turbocharger compressors led to the present investigation. A dedicated test rig was built to experimentally study two different production turbocharger compressors of equal nominal dimensions: the ‘C288’ and ‘C117’, as used in the CAT C9.3 ACERT™ industrial engine. These compressors were tested under steady conditions – providing a performance map extending beyond the supplier surge line, and transient conditions – providing measurements of the compression system undergoing surge. The shape of the steady speed lines was found to be much flatter in the C117 compressor, which drives a more gradual transition from mild to deep surge than the C288 compressor. The compressors also exhibited contrasting frequencies of mild surge (9.6 Hz and 5.1 Hz for the C288 and C117, respectively). A surge model was developed in the Dynasty simulation tool. Alterations to existing code extrapolated compressor performance to the stalled and reverse flow regions, and implemented a more accurate representation of energy flow and use of upstream conditions for map correction during reverse flow. These improvements allowed the simulation to match the measured frequencies and amplitudes of data from literature and the present experimental campaign. For example, mild surge simulations in Dynasty predicted a frequency of 9.7Hz, very close to the measured value of 9.6Hz. Deep surge periods were simulated within 0.4% and mass flow amplitude within 9.2%, of their measured values, demonstrating excellent agreement and proving the model’s predictive capabilities. A novel compressor shaft torque sensor was developed to address the lack of knowledge of how compressor torque varies during surge events, in literature and industry. The sensor was used to measure torque during mild and deep surge transients, providing data that is the first of its kind. The experimental results and numerical methodologies reported herein provide improvements to the understanding and prediction of surge behaviour, and pave the way for additional research in this field.
Content Version: Open Access
Issue Date: Dec-2019
Date Awarded: Sep-2020
URI: http://hdl.handle.net/10044/1/99659
DOI: https://doi.org/10.25560/99659
Copyright Statement: Creative Commons Attribution NonCommercial Licence
Supervisor: Costall, Aaron
Martinez-Botas, Ricardo
Sponsor/Funder: Caterpillar Inc
Department: Mechanical Engineering
Publisher: Imperial College London
Qualification Level: Doctoral
Qualification Name: Doctor of Philosophy (PhD)
Appears in Collections:Mechanical Engineering PhD theses



This item is licensed under a Creative Commons License Creative Commons