<p>The purpose of this research program was to investigate the void behaviour in a vertical conduit containing a flowing mixture of air and water.</p> <p>The local void fraction was measured by a neutron attenuation technique which was developed for this study. A neutron beam extracted from McMaster's 2MW Swimming Pool Nuclear Reactor was conditioned to provide a collimated beam of thermal neutrons with low gamma and fast neutron back-ground. An atmospheric vertical upward cocurrent air-water loop was constructed in front of the beam port.</p> <p>A series of experiments was performed to obtain information on the steady-state radial void profile in the bubbly and annular flow regimes for a 3/4 in. I.D. test section with gas and liquid volumetric fluxes up to about 32 m/sec and 80 cm/sec, respectively. The dynamic behaviour of the system to a perturbation in gas flowrate was measured and compared to that predicted by the void propagation equation.</p> <p>The adequacy of the netron attenuation method for radial void profile measurements in two-phase flow was ascertained. This was achieved by integrating the measured radial void profile and comparing the resulting cross-sectional average void fraction with that obtained by a trapping method. The neutron attenuation method was also shown to be an excellent tool for flow regime recognition and for testing the flow development. The effect of the statistical fluctuations of the neutron source on its capability for void fraction measurement was examined and a counting strategy was recommended to reduce the resulting bias.</p> <p>From the steady-state measurements of the radial void profiles and the gas and liquid flowrates, estimates of important flow parameters were obtained for the fully-developed bubbly and annular flow regimes. These estimates indicated that the flow parameters remain essentially constant for a specific flow regime.</p> <p>To reduce the experimental effort considerably, an alternative method was presented which enables the designer to evaluate these parameters by performing a single experiment where the radial void profile is measured. Based on the momentum equation and Prandtl's turbulent mixing length hypothesis, a general form of the shear stress radical distribution was derived and intergrated to yield the radial profile of the mixture velocity. By combining the radial profiles of the mixture velocity and void fraction, the values of the flow parameters were predicted. The proposed method indicated systematic variations of the flow parameters with the gas and liquid flowrates which were not consistent with observations. The estimated flow parameters, however, were shown to be adequate to predict the average void fraction without introducing appreciable errors.</p> <p>The results of the current experimental observations and analysis were compared with selected models and correlations which have been used by many designers to describe two-phase flow behaviour. This was due to evaluate the validity of the assumptions made in deriving these models and correlations. In general, this comparison indicated that the present experimental results and model predictions were consistent with expectations and previous observations and predictions. However, many of the published models and correlations were found to yield misleading results and conclusions which were attributed to the simplifying assumptions included in their derivation.</p> <p>The void propagation equation was applied to predict the transient response of void fraction to perturbations in the gas flowrate in the adiabatic gas-liquid flow system employed. To solve this equation, values of the distribution parameter are required under changing flow conditions. It is usually assumed that the parameter values obtained for steady-state, fully-developed flow are valid under transient developing-slow situations. That is to say that the steady-state distribution parameter pertaining to the local average voidage condition applies in the transient system. Since this parameter cannot be obtained from steady-state experiments under all voidage conditions experienced during the transient, a continuous functional relationship was assumed in order to interpolate over those regions where the parameter cannot be obtained.</p> <p>An analysis was performed to examine the sensitivity of the predicted void transient response to errors in the distribution parameter. This analysis showed that a comparison between predicted and observed void transient responses should permit the evaluation of the validity of the assumptions made concerning the distribution parameter and its interpolating relationship. A satisfactory agreement between these responses was obtained in bubbly flow. This indicated that the steady-state values of the distribution parameter for fully-developed flow and the assumed interpolating relationship can be used to predict the void transient response in the considered experimental apparatus over the flow conditions studied herein.</p> / Doctor of Philosophy (PhD)
| Identifer | oai:union.ndltd.org:mcmaster.ca/oai:macsphere.mcmaster.ca:11375/11401 |
| Date | 10 1900 |
| Creators | Younis, Hosny Mohamed |
| Contributors | Hoffman, T. W., Harms, A. A., Chemical Engineering |
| Source Sets | McMaster University |
| Detected Language | English |
| Type | thesis |
Page generated in 0.002 seconds