Steam condensation in flattened-tube air-cooled condensers

Air-cooled condensers for power plants are an alternative to closed-cycle wet cooling with a condenser and cooling tower pair. Air-cooled condensers cool the process steam by forced-convection of air, replacing the water evaporation found in traditional power-plant designs. Air-cooled condensers are rising in prominence because they provide utility companies with additional freedoms – to build power plants away from large water sources and to avoid a lengthy environmental permitting process. In addition, the environmental benefits of reducing power-plant water consumption are significant. However, the lower thermal performance of air-cooling in comparison to water-cooling makes the cost of air-cooled condensers prohibitive in many cases. In order reduce this performance gap, the current designs of air-cooled condensers must be significantly improved. However, to improve the design, understanding of the condenser physics through experimental investigation is needed. As of yet, there are no experimental results in the open literature for steam-side performance of the most-common condenser design - the flattened-tube air-cooled condenser. This thesis provides the experimental results and analysis to address this deficiency. The experiments contained herein provide visualization along with measurement of void fraction, pressure drop, capacity and heat transfer coefficient for flattened-tube air-cooled steam condensers with inner dimensions of 216 mm height x 16 mm width. Two lengths of tube are investigated: 10.7 m and 5.7 m. The tubes are steel with brazed wavy aluminum fins of 200 mm length x 19 mm height. Air is in cross-flow to the condensing steam. Typical condenser designs contain about 80% of the tubes in co-current configuration, with both steam vapor and liquid flowing downwards. The remaining 20% of tubes are in counter-flow configuration, with vapor flowing upwards and condensate flowing downwards. This thesis investigates only the co-current downward-flowing tubes. In addition, the effect of downward inclination angle of the tubes is investigated, with the condensers mounted on a hinged truss in the test facility to enable lifting to inclination angles from 0o to 75o. Initial experiments are performed on the tube with 10.7 m length. The condenser tube is cut in half lengthwise and covered with a polycarbonate window to perform diabatic visualization simultaneously with the heat transfer and pressure drop measurements. The effect of tube inclination on flow regimes, void fraction, pressure drop, capacity and heat transfer coefficient (HTC) is evaluated. The flow regime is found to be stratified for almost all conditions, with stratified-wavy flow observed near the inlet of the horizontal tube and near the outlet of tubes inclined 60o or greater. Increasing downward inclination angle is found to increase the drainage of condensate from the tube, thereby increasing the average void fraction. The greatest increase in void fraction is seen near the tube outlet. Increasing inclination is also seen to decrease the pressure drop in the tube. This is the result of gravitational pressure recovery, as well as decreased vapor velocity from the increased void fraction. Increasing inclination also increases capacity, as the improved drainage of condensate reduces the condenser thermal resistance. Following these initial results in the half-tube, a full (un-cut) tube is tested in the same experimental facility. This tube has a shorter length of 5.7 m and is able to be tested at lower condensing pressures. For this set of experiments, adiabatic visualization sections are designed and installed at the condenser inlet and outlet to provide identification of flow regimes and measurement of void fraction. The tube inclination is varied from 0o to 49o downwards, and the condensation pressure is varied from 60 kPa to 105 kPa. Similar results for flow regimes and void fraction are found as in the 10.7 m tube, with annular flow at the tube inlet and stratified flow at the tube outlet. Increasing tube inclination is found to decrease the depth of the condensate river. This decreasing depth is found to decrease frictional pressure drop and increase condenser heat transfer coefficient. The increase in heat transfer coefficient due to inclination is found only near the tube outlet, and is smaller than the increase measured in the 10.7 m tube. Overall, the results show that the reduced tube length leads to less condensate build-up and improved thermal performance. From these experimental results, a thermo-hydraulic model for void fraction and capacity is developed. The model uses open-channel-flow theory to predict the depth and velocity of the condensate in the stratified layer at the tube bottom. The thermal model calculates local HTC on the air and steam sides and can provide a local description of heat flux in the condenser. The thermal model is validated by experiments in both the 10.7 m half tube and the 5.7 m full tube. The model predicts 86% of the experimental results for condensate river depth to within 20% accuracy and 98% of the experimental results for capacity within 5% accuracy. Using the model, alternative airflow profiles that match non-uniformities on the air and steam sides of the condenser are proposed and tested experimentally. The first profile reverses airflow direction and yields an increase in condenser capacity of 3.5%. The second profile increases airflow near the steam inlet and reduces airflow near the steam outlet. This yields a 3.1% increase in capacity as well as a reduction in pressure drop (for a constant capacity). Finally, to more accurately measure the steam-side heat transfer coefficient, a new condenser test section is designed. This test section replaces the air-cooled fins with a single pass of cooling water. The design maintains equivalent operating conditions with the air-cooled condenser while providing accurate local heat transfer coefficient determination. In particular, the test section maintains the non-uniform heat flux and wall temperature of the air-cooled section. The results show that two distinct heat transfer regions can be defined: the stratified condensate layer flowing axially at the tube bottom, and the condensing-film region along the tube side-walls and top. The heat transfer coefficient in the stratified layer depends mostly on the condensate depth, and the heat transfer coefficient in the condensing-film region depends mostly on the wall-steam temperature difference. The heat transfer coefficient in the stratified layer is much lower in magnitude (up to 30 times) than that in the condensing-film region. The heat transfer coefficient in the stratified region is under-predicted by the correlations of Rosson and Myers (1965) and Dobson and Chato (1998). In the condensing-film region, the heat transfer coefficient can be accurately predicted by the theory of Nusselt (1916) when considering the local wall-steam temperature difference. The experimental results are not accurately predicted by existing correlations for heat transfer coefficient during condensation, so a new correlation is developed that is a perimeter-weighted average of the heat transfer coefficient in the stratified condensate and condensing-film regions. The new correlation predicts 77% of the data within 20% accuracy, with a mean absolute percent error of 14%