This thesis describes the development of experimental and predictive techniques for
analysing boiling heat transfer for internal combustion engine cooling systems.
Environmental and packaging requirements are placing increasing demands on the
emission and power density characteristics of internal combustion engines. To meet
these demands, boiling heat transfer is proposed as a means of achieving higher heat
transfer rates than is possible with conventional forced convection.
A novel experimental method developed using electrically heated indium tin oxide
(ITO) coated glass in conjunction with a bubble image analysis system allowed new
boiling data to be obtained. This allowed the study of boiling from a new perspective
with higher data analysis rates than was possible in previous reported work. It also
enabled, for the first time, direct, non-contact measurement of bubble/liquid void
The data from these experiments was compared with predictions from a
computational fluid dynamics (CFO) package known as CFX-4. The predictions for
the percentage of the heated surface covered in vapour were shown to be strongly
dependent on the bubble diameter and nucleation site density submodels. These
values were in turn demonstrated to be dependent on the surface finish of the heated
ITO coated glass was successfully implemented for the study of boiling heat transfer.
The data from this work allowed a new evaluation of the predictive abilities of the
CFO boiling models. The standard default bubble diameter and site density submodel
equations, whose lack of universal applicability has been demonstrated, were
replaced with new empirically based equations that provided more accurate
predictions. As a result, a new design methodology for practical application of
nucleate boiling to rc engme cooling galleries was proposed. This method
introduced an experimental stage early in the design process to obtain bubble
diameter and site density data for when boiling occurs on the surface to be used in
the finished product. These data provide empirically based submodels for use in CFO
subroutines. These ensure that more accurate results can be obtained than would be
possible with the standard default submodel equations.
A Doctoral Thesis. Submitted in partial fulfilment of the requirements for the award of Doctor of Philosophy of Loughborough University.