An empirical approach to predicting heat transfer within single- and twin-skin automotive exhaust systems

Christopher D Bannister, Christian J Brace, J Taylor, T Brooks, N Fraser

Research output: Contribution to journalArticle

2 Citations (Scopus)
136 Downloads (Pure)

Abstract

This paper describes the further development of an exhaust system model based on the experimental characterisation of heat transfer in a series of different pipe sections. Building on previous work published in this journal by the authors, this study was undertaken to improve the operating range, accuracy and usability of the original model as well as introducing the ability to model twin skin exhaust sections with an air gap.
Convective heat transfer relationships for nine stainless steel exhaust bend sections of varying wall thicknesses and radiuses were experimentally characterised over a range of steady state conditions. In each case a correlation between observed Reynolds number (Re) and Nusselt number (Nu) was developed. Based on measured experimental data, a generic model was built using Matlab/Simulink capable of predicting the relationship between Nusselt number and Reynolds number for previously unseen pipe geometries falling within the experimental design range. To further develop the usefulness of the model, fifteen twin skin test sections, intended to represent a range of geometries applicable to production automotive gasoline exhaust systems, were also fabricated and characterised. Within the model, both skins of each pipe section were split into five axial and radial elements with the inner and outer skins linked via the modelling of free convection and radiation between them.
The predicted Reynolds-Nusselt relationships for each bend section and twin skin configuration were validated using transient experimental data over a portion of the US06 drive cycle. The final model demonstrated improved accuracy of exhaust gas temperature predictions, compared with previous model iterations, with typical errors of less than 1% and a mean error over the US06 cycle of +0.2%.
Original languageEnglish
Pages (from-to)913-929
Number of pages17
JournalProceedings of the Institution of Mechanical Engineers, Part D: Journal of Automobile Engineering
Volume225
Issue number7
Early online date10 May 2011
DOIs
Publication statusPublished - Jul 2011

Keywords

  • heat transfer
  • modelling
  • exhaust system

Fingerprint Dive into the research topics of 'An empirical approach to predicting heat transfer within single- and twin-skin automotive exhaust systems'. Together they form a unique fingerprint.

  • Cite this