AbstractThe performance of an industrial furnace or boiler is dictated by three interrelated processes, namely: combustion, furnace aerodynamics, and heat transfer. Thus, the flow behaviour of the gases in a furnace chamber can have a significant influence on the combustion and heat transfer processes and hence the performance of the furnace.
The rate of combustion depends on the rate of mixing between the fuel, air and combustion products and this can be enhanced either by increasing the jet momentum or by imparting a swirling motion to the burner fluids. Thus, the application of swirl is very common and there is also some evidence that it can lead to an increase in heat transfer to the load and walls of the enclosure. In most furnace applications radiation accounts for 90 - 95 % of the total heat transfer and convection is relatively small. However, there are applications where flame temperatures and emissivities are relatively low and convection becomes significant and accounts for up to approximately 50 % of the overall heat transfer.
At the present time designers of industrial furnaces or boilers are often restricted to using a single value for the convective heat transfer based on empirical data for fully-developed duct flow since there is a lack of more reliable data on convective heat transfer.
This PhD project aims to provide convective heat transfer data which can be applied in industrial furnace or boiler applications. The flow regimes developed by a burner firing into a furnace chamber are similar to those encountered downstream of a sudden expansion in a duct. The literature survey highlighted the lack of information on the effects of swirl on convective heat transfer in this geometry so that this topic is the main subject of the present thesis. Swirl was imparted with the aid of tangential inlets and the effect of the changes in the degree of swirl (including the zero swirl case), inlet to outlet duct diameter ratios, and inlet Reynolds numbers on the convective heat transfer were
Laser Doppler Anemometry was initially employed to study the velocity field in the outlet pipe, particularly in the non-swirling case. Liquid crystal thermometry was utilised to obtain the variation in convective heat transfer coefficients on the duct wall. The technique used in the project was an improvement on previously published methods in that the liquid crystal images were stored directly in the computer for subsequent analysis.
Previously, these images have been recorded on video with the possibility of distortion in the subsequent image analysis. The convective heat transfer data were normalised using the fully-developed duct flow value as a base. It was found that the normalised peak convective heat transfer increases with decreasing Reynolds number, increasing diameter ratio and increasing degree of swirl in most cases. However, for large expansions (i.e. relatively small inlet diameters) a small increase in the degree of swirl initially resulted in a decrease in the peak convective heat transfer. The application of higher rates of swirl, however, subsequently results in significant enhancement in heat transfer.
Furthermore, swirl only increased the average heat transfer over a relatively long outlet pipe for a large diameter ratio, whereas at smaller diameter ratios the average heat transfer decreased with swirl in comparison to the no swirl value, even for short outlet pipe lengths.
The problem was also studied numerically using a computational fluid dynamics package, CFX-F3D. A differential Reynolds stress turbulence model was employed and measured tangential velocities were used as input for the swirling flows. The predicted convective heat transfer coefficients were generally in good agreement with the measured data.
|Date of Award||1997|
|Supervisor||John Ward (Supervisor)|