SUBJECT: M.S. Thesis Presentation
   
BY: James Golden
   
TIME: Thursday, June 21, 2012, 9:30 a.m.
   
PLACE: Love Building, 210
   
TITLE: Ammonia-Water Desorption in Flooded Columns
   
COMMITTEE: Dr. Srinivas Garimella, Chair (ME)
Dr. S. Mostafa Ghiaasiaan (ME)
Dr. Sheldon Jeter (ME)
 

SUMMARY

Refrigeration systems employing the NH3-H2O absorption cycle provide cooling using a thermal energy input. This cycle relies on the zeotropic nature of the refrigerant – absorbent pair: because of the difference in boiling temperatures between NH3 and H2O, they can be separated through selective boiling in the desorber. Desorbers with counter-current flow of the solution and generated vapor enable efficient heat and mass transfer between the two phases, reducing the absorbent content in the generated vapor. Flow visualization experiments at temperatures, concentrations and pressures representative of operating conditions are necessary to understand the heat and mass transfer processes and flow regime characteristics within the component. In this study, a Flooded Column desorber, which accomplishes desorption of the refrigerant vapor through a combination of falling-film and pool boiling, was fabricated and tested. Refrigerant-rich solution enters the top of the component and fills a column, which is heated by an adjacent heated microchannel array. The vapor generated within the component is removed from the top of the component, while the dilute solution drains from the bottom. Flow visualization experiments showed that the Flooded Column desorber operated most stably in a partially flooded condition, with a pool-boiling region below a falling-film region. Heat transfer coefficients were calculated from the data for the pool-boiling region, and were compared with the predictions of several mixture pool-boiling correlations from the literature. The correlations from the literature were in general unable to predict the data from this study adequately. Modifications to existing mixture boiling correlations are suggested based on the findings of this study. The resulting modified correlation predicts 33 of the 35 data points from this study within ±40%, with an average absolute error of 19%.