Abstract
This dissertation is comprised of two parts which consider separate aspects of gas/liquid two-phase flow: flow in horizontal circular pipes and flow in vertical corrugated channels typical of Plate-and-Frame Heat Exchangers (PHEs). A review of the literature has revealed the lack of any formal analyses of the performance of phenomenological methods for the prediction of pressure drop during gas/liquid two-phase flow in cylindrical pipe flow. A number of models describing various horizontal flow patterns have been evaluated against a large experimental data bank using a logarithmic statistical analysis technique. Data are categorised by the Taitel & Dukler (1976) flow pattern map. It is shown that the predictions of better phenomenological models are equal to those of empirical methods, while the probability density functions are less sensitive to changes in fluid system. A qualitative approach is developed to observe variation in error as a function of flow pattern. Error is shown to be a function of position on the flow pattern map for a number of phenomenological models, while the relationship is markedly weaker for empirical methods. It is observed that flow pattern transition influences prediction performance, particularly with the onset of intermittent flow phenomena. Two composite methods consisting only of phenomenological models are defined using both the quantitative statistical analysis results and the qualitative flow pattern based observations. Method 3 generates predictions that are more reliable over the entire range of the data bank than all empirical methods included in the study. It achieves a 10% reduction in scatter (Sln) compared to the best empirical method, that of Bandel (1973). It is also shown to be less sensitive to changes in fluid system. Measurements of pressure drop during adiabatic single-phase and two-phase flows in PHEs are also presented. Liquid flows of water and CMC solutions of varying concentration were combined with air to provide the two-phase flow. Two-phase pressure drop multipliers are presented. While the relationship between the two-phase multiplier for the overall test section and the Lockhart-Martinelli parameter varies as a function of chevron angle a common function is defined for the corrugated section; a curve fit of this function is presented. Flow pattern maps are defined for PHEs. They include regions of bubble, chum, film and partial film flows. Observed interfacial structures show similarities with those seen in other low hydraulic diameter systems. Flow pattern transitions for the 30/30 and 30/60 configurations are similar, indicating that the lower angle is dominant in the mixed B channel. This observation parallels that of Focke & Knibbe (1986) for single phase flow paths. The map obtained for the channel with 60/60 chevron geometry is significantly different, indicating a clear geometry effect.