L. C.S. Mesquita, D. Thomey, S. J. Harrison.

Department of Mechanical and Materials Engineering, McLaughlin Hall, Queen’s University, Kingston, ON. Canada, K7L 3N6. Email :mesquita@me. queensu. ca

INTRODUCTION

In the last few years there has been renewed interest in solar driven air-conditioning

[1] . Some of the work have been focused in desiccant cooling systems. Such systems have the advantage of improved humidity control, particularly in applications with high ventilation rates [2]. Most of the systems already developed employ solid desiccants, with relatively high regeneration temperatures. One alternative is the use of liquid-desiccant systems. In these systems, lower regeneration temperatures can be employed, allowing for a more efficient use of heat from low temperature sources, e. g., flat plate solar collectors [3]. Another advantage of liquid-desiccant systems is the potential of using the desiccant solution for energy storage.

The main components in a liquid-desiccant air-conditioning system are the dehumidifier and the regenerator. Many different technologies have been developed for these two components. For the dehumidifier, the most common technology employed today is the packed bed. However, packed beds must work with high

desiccant flow rates, in order to achieve good dehumidification levels without internal cooling. Higher desiccant flow rates imply on small changes in the concentration of the desiccant solution during the process. This, and the higher level of heat dumping from the regenerated solution that follows higher flow rates, reduce the coefficient of performance of the liquid-desiccant cycle. One option that allows lower flow rates is the use of internally cooled dehumidifiers [4,5]. Figure 1 presents the schematics for one channel of a internally cooled dehumidifier, wich is composed of several of these channels stacked together.

In the present work, mathematical and numerical models were developed for internally cooled dehumidifiers, using three different approaches. The first approach uses heat and mass transfer correlations. The second one numerically solves the differential equations for energy and species for a constant thickness film, using the finite-difference method. The third approach introduces a variable film thickness. All approaches assume fully developed laminar flow for the liquid and air streams.