Experimental data are reported for boiling water and a glycerol-water mixture at a free surface pressure of 50 mbar absolute on the shell-side of a thin slice model of an industrial boiler. The boiler test section was 1 m high, 0.75 m wide, 98 mm long and contained 36 electrically heated, horizontal tubes that were 28.5 mm in diameter. The design of the boiler ensured that the tubes were submerged in a liquid pool. The height of the liquid pool was set to 2 m, submerging the top of the tube bundle in 1.6 m of liquid. The heat flux was varied within the range 10-65 kW/m(2). A near-symmetrical half of the tube bundle contained wall thermocouples. An additional 29 thermocouples were located throughout the liquid pool.For both fluids, the liquid temperature in the pool was found to be reasonably uniform and controlled by the pressure at the free surface. This led to subcoolings of up to 31 K on the tube surfaces. The reasonably uniform pool temperature suggests that the liquid re-circulates within it.For water, boiling was initiated in the heat flux range 25-40 kW/m(2), whereas the glycerol-water mixture initiated boiled within the range 10-25 kW/m(2). Below these heat flux ranges, both fluids were in natural convection, with the measured wall superheats in reasonable agreement with predictions from a correlation available in the open literature. The difference in the fluids' boiling onset resulted from the natural convection, heat-transfer coefficients of the glycerol-water mixture being lower than that for water. The boiling wall superheats for water were reasonably well predicted by a correlation available in the open literature.Boiling glycerol-water mixture data, taken at atmospheric pressure and available in the open literature, was used to identify methods for correcting pure fluid boiling heat-transfer coefficients for mixture effects. Mixture boiling superheats were reasonably well predicted by some of these methods. The method that worked best at atmospheric pressure did not work best at low pressure. A method is identified that is reasonable at atmospheric and low pressures. (C) 2016 Elsevier Ltd. All rights reserved.

Experimental data are reported for water boiling at pressures of 850 and 50 mbar absolute on the shellside of a model industrial boiler slice. The boiler test section was 1 m high, 0.75 m wide and contained 36 electrically heated tubes. The tubes were 28.5 mm in diameter and 98 mm long. The design of the boiler ensured that the tubes were submerged in a liquid pool. The height of the liquid pool could be varied. The pool height was set to approximately 0.8 m for the tests carried out at a pressure of 850 mbar, submerging the top of the tube bundle by about 200 mm. Two pool heights were used in the tests carried out at a pressure of 50 mbar, one at approximately 0.8 m and another at approximately 2 m. The later submerged the top of the tube bundle by about 1.6 m. The heat flux was varied within the range 10-70 kW/m2. A near-symmetrical half of the tube bundle contained wall thermocouples. An additional 29 thermocouples were located throughout the liquid pool.The liquid temperature in the pool was found to be reasonably uniform and controlled by the pressure at the free surface. This led to a small amount of subcooling at a pressure of 850 mbar, up to 3 K, and a significant amount of subcooling at a pressure of 50 mbar, up to 16 K for the smaller pool height and up to 31 K for the larger pool height. The reasonably uniform pool temperature suggests that the liquid re-circulates within it.Boiling was found to occur at all heat fluxes at a pressure of 850 mbar, with the measured heat-transfer coefficients shown to be in broad agreement with nucleate boiling correlations available in the open literature. However, they were also consistent with a flow boiling process involving natural convection and nucleation, where the convection was driven by variations in liquid temperature on the walls of the tubes. This natural convection relies on an interaction between the tubes that produces mass fluxes in the range 46-87 kg/m(2) s, based on the approach area to the tube bundle. Boiling occurred only at the higher heat fluxes during the low level tests at a pressure of 50 mbar, with interactive natural convection being the dominant heat-transfer mechanism. The mass fluxes produced were in the range 28-70 kg/m(2) s. Boiling also occurred only at the higher heat fluxes during the high level tests at a pressure of 50 mbar. However, the convective heat transfer was more compatible with little interaction between the tubes, although some evidence suggests that the evaporator oscillates between interactive and isolated tube behaviour. (C) 2015 Elsevier Ltd. All rights reserved.

The two-fluid model is applied to a thin sliced kettle reboiler. The tube bundle is treated as a porous medium in which the drag coefficient and tube-wall force are deduced from the empirically-based, one-dimensional model. Methods available in the open literature are used in the two-phase pool surrounding the tube bundle. The predictions are verified by comparing them with experimental data and models available in the open literature.The boundary condition applied at the free surface of the pool is found to be crucial in determining the flow pattern within it. When only liquid re-enters through the boundary an all-liquid pool results. Comparison with the experimental evidence suggests that this boundary condition corresponds to bubbly flow within the tube bundle. Allowing a predominantly vapour re-entry produces a two-phase pool that IS consistent with intermittent flow in the tube bundle. When the appropriate boundary condition is applied, the two-fluid model predictions are shown to reproduce the visual records and pressure drop measurements reasonably accurately. (C) 2011 Elsevier Ltd. All rights reserved.

A particle image velocimetry (PIV) autocorrelation technique was applied to produce whole field vector maps of flow beneath and to the side of a 17 row×l7 column kettle reboiler thin slice rig, boiling pentane at atmospheric pressure. The flow proved to be time dependent. The average values of the mass flowrates of the recirculating liquid were evaluated at various positions and compared to predictions of models of the flow which reproduced the measured pressure drops [B.M. Burnside, K.M. Miller, D.A. McNeil T. Bruce, Heat transfer coefficient distributions in an experimental kettle reboiler thin slice, Trans. IChemE 79A (2001) 445–452]. The use of the results as a platform for 2D numerical modelling of the flow is emphasised.

The paper describes tests boiling R113 at atmospheric pressure in upward flow over a 17 row column of square pitched electrically heated tubes. Uniform heat fluxes of 10-65 kW m(-2) and maximum Reynolds numbers, Re-max, between 7800 and 27000 were used. The data at the lower heat fluxes of 10 and 20 kW m(-2) consistently exhibited a linear increase of heat transfer coefficient, h, with quality, the slope lower at 20 kW m(-2), so that h at 10 kW m(-2) exceeded that at 20 kW m(-2) and both reached h at 40 kW m(-2) for the maximum quality tested. At the higher heat fluxes a nucleate boiling controlled region was observed at low quality, with h equal to the value at the same heat flux as observed with an isolated tube in a pool, followed by a rise in h at higher quality. The results were compared with the work of other researchers. Heat transfer coefficients for q >= 40 kW m(-2) were predicted to an average r.m.s. deviation of 7%, using the asymptotic flow boiling model. However, the sensitivity of h to change of quality and flowrate predicted was much lower than measured and in some cases exhibited opposite trends.

Pressure drop measurements were carried out in a test condenser of staggered configuration with p/D = 1.33. The test conditions were: steam inlet pressure, 50 mbar, inlet velocity 10-30 ms(-1) and steam-to-cooling water temperature difference 5-15 K. This corresponded to a mean bundle Re-v,Re-max between 1000 and 7000, suction parameter range 0.3-2.8 and heat flux densities at the tube outside wall up to 79 kW m(-2). In the range 2800 < Re-v,Re-max < 6400 increase in condensation rate progressively reduced the pressure drop coefficient to a minimum of 27-30% of the corresponding dry flow at Re,,m,x = 2840 and 5320 as predicted by ESDU 74040. The results were compared with previous investigations and condensing pressure drops found comparable with Nicol et al. [Proc. 7th Int. Heat Transfer Conf. 5 (1982) 133-138] but the Fujii et al. [Int. J. Heat Mass Transfer 15 (1972) 247-260] pressure drops were always found to be lower by up to 52% than the present data. The flow patterns associated with the effect of condensation on pressure drop are discussed.The results reported here suggest that considerable errors in pressure drop, with consequent errors in heat transfer distribution, can result from ignoring the effect of suction as is currently done in condenser design and prediction. (C) 2004 Elsevier Ltd. All rights reserved.

An experimental study has been undertaken into the enhancement obtained in the heat-transfer coefficients when HIGHFLUX tubes are used in preference to plain tubes while boiling pentane. The study involved two experimental facilities, a single-tube pool boiler and a 241 tube, 17 row by 17 column, thin slice kettle reboiler. The pool boiling results show that the HIGHFLUX tubes produce heat-transfer coefficients that are up to five times larger than their plain tube counterparts. In flow boiling the enhancement is 3–6 times. In both cases, HIGHFLUX tube performance is shown to deteriorate when small degrees of subcooling are present in the liquid. The deterioration still leaves the HIGHFLUX tubes with a significantly higher heat-transfer coefficient than the plain tubes. Existing flow boiling design methodologies are shown to produce performance characteristics that HIGHFLUX tubes do not follow.

Abstract A purpose-built test facility has been constructed and used to produce data for filmwise condensation from steam, and steam-air mixtures, flowing downwards across a 15 row by 5 column bundle of tubes. Data were obtained at conditions typical of those found in the UK electricity generating industry. Steam was supplied at pressures of 50, 75 and 100 mbar, at velocities of 10,20 and 33 m/s and with air concentrations of 0 and 10000 ppm. Steam to cooling water temperature differences of 5, 10 and 15 K were used to generate heat fluxes of up to 90 kW/m2. The data were used to investigate the application to tube bundle design. It is shown that the selection of an appropriate steam velocity allows good agreement between some shell-side heat transfer correlations and the experimental data, and that existing correlations for inundation and air-concentration effects are sufficient.

Pressure drop measurements in a 15 row staggered configuration condenser, p/D = 1.33, are described. Heat flux densities of up to 90 kW(.)m(-2) were used. The results are compared with the those of another investigation using a condenser with much less closely packed tubes. At relatively low suction, pressure drop in the closely packed condenser could be predicted by conventional single phase correlations. At higher values of suction the condensing pressure toss was considerably lower in this condenser than the single phase value. In the more loosely packed condenser this difference occurred at all values of suction. It is concluded that more experimental and theoretical work is required to explain these effects in view of their importance in turbine condenser design. (C) 2001 Editions scientifiques et medicales Elsevier SAS.

Data have been produced for filmwise condensation of steam, and steam–air mixtures, flowing downwards across two tube bank, a 15 row, in-line bank containing 75 tubes and a 15 row staggered bank containing 82 tubes. Both banks were tested at conditions typical of those found in the UK electricity generating industry. Steam was supplied at pressures of 50, 75 and 100 mbar, at velocities of 10, 20 and 33 m/s and with air concentrations of 0 and 10,000 ppm. Steam to cooling water temperature differences of 5, 10 and 15 K were used to generate heat fluxes of up to 90kW/m2. The data and a mathematical model were used to investigate the effect of geometry difference on the heat transfer and pressure difference characteristics. For this particular configuration, the staggered and the in-line tube banks gave the same performance. This is not consistent with similar studies done by other researchers and indicates that tube spacing is important.

Pressure drop and heat transfer coefficient measurements have been made in a 241 tube bundle kettle reboiler thin slice rig, boiling pentane at atmospheric pressure. The effects of liquid/vapour separation problems in the shell are described and discussed. A conventional 1-D recirculation model predicted the data taken at a uniform heat flux of l0 kW m−2. Another recirculation model was developed and compared to the data at 50 kW m−2where the flow in the bundle was markedly two dimensional. This model matched the pressure drop data in the bundle apart from that in the outside columns which were judged to have been affected by adjacent rapidly rising 2-phase flow in the shell. The model was adjusted to eliminate this defect and making use of two flow boiling models reproduced the heat transfer data mainly within experimental error. The consequence of these findings for the application of data from thin sliced rigs to design of full scale kettle reboilers and horizontal recirculation shellside evaporators is discussed.

Pressure drop measurements in a 15-row steam condenser configured with in-line tubes, p/D < 1.33, are described. Pressures from 50 to 100 mb and Remax in the range 1,000-18,000 were imposed. It is shown that pressure loss coefficients for the bundle and for two-row pairs were lower than predictions for equivalent single-phase tests except near the bottom of the bundle. There was some evidence that increase in suction parameter increased this effect in the top rows. Taken together with previous investigations [3, 4], a falling trend of suction effect is evident the more closely packed the tubes are. The discrepancy between these findings and the results of simulation experiments [7] is noted.

Application of dropwise condensation to utility turbine condensers is investigated by comparing the thermal performance of dropwise and filmwise bundles at industrially relevant conditions. Steam and steam-air mixtures were condensed on bundles of in-line, titanium tubes. The row-by-row heat transfer coefficients are presented against bundle position. They show the expected behavior for filmwise condensation but demonstrate a different one for dropwise. In air-free steam, the dropwise heat transfer coefficients are much larger and do not vary significantly with bundle position. In air-steam mixtures the dropwise values decrease similarly to their filmwise equivalents. The findings are in accord with those found for other geometries. The findings indicate that significant reductions in condenser size can be obtained if permanent dropwise condensation can be produced at industrially relevant conditions.

Pressure drop measurements in boiling R113 at 1.013 bar in upwards flow over a horizontal tube bundle are described. Uniform heat flux densities from 5 to 65 kW/m(2) and mass velocities in the range 200 to 700 kg/m(2)s were imposed. The measured data is compared to three current prediction methods all of which generally overestimated it. The predictions of the best model were on average only 1% higher than the measurements with a standard deviation of +/- 9%. The discussion speculates that there is a heat flux dependence of pressure drop, which increases with mass flux.

A computer simulation of dropwise condensation of steam on a 240 x 240 mu m surface with 60,000 randomly spaced nucleation sites is described. The maximum drop radius achieved was 3.9 mu m, 0.21 ms after the start of condensation. Uniform radii drop generations noticed by Rose and Glicksman (1973) had not developed completely by 0.21 ms, although predicted to do so. The characteristic profile on the drop size histogram predicted by Tanaka (1979) extends to drop sizes smaller than have been observed in experiments. A peak heat transfer coefficient of just over 2 MW/m(2)K was obtained, about twice the value measured by Tanasawa et al. (1978) immediately after the condensing surface had been wiped. (C) 1999 Elsevier Science Ltd. All rights reserved.