All liquid-desiccant air conditioners require a regenerator--the component that concentrates the dilute desiccant by removing excess water.  The simplest regenerator is a single-stage scavenging-air regenerator in which the desiccant is heated to evaporate water into a scavenging air stream.  When regenerating lithium chloride to a high concentration, this regenerator will have a thermal COP of between 0.6 and 0.8 (when it is used with an air-to-air heat exchanger that preheats the scavenging air).

It is well known that multiple-effect boilers can be used to increase the efficiency of separation processes such as the regeneration of desiccant.  Their application to removing water from the lithium-bromine brine in  absorption chillers has increased the COP of these cooling systems from about 0.6 to 1.0.

Although a double-effect boiler can be used to regenerate desiccant, a simpler approach, which achieves almost the same COP, is the two-stage 1½-effect regenerator.  As shown in the preceding figure, heat is applied to a first-stage atmospheric-pressure boiler.  The steam that leaves this boiler has a saturation temperature of 212^oF.  It is used to heat a secondary hot-water loop, the heating occurring in the direct-contact condenser that is located between the boiler and the scavenging-air (S-A) regenerator.  This secondary hot-water loop is then the heat source for the second-stage Scavenging-Air Regenerator.

In September 2005, we completed a proof-of-concept model of the 1½-effect regenerator with funds provided by the U.S. Department of Energy through their Invention and Innovation program.  As shown in the neighboring picture, the first-stage boiler is a modified, coiled-tube hot water boiler.   A spin separator at the exit of this boiler effectively removes all desiccant droplets from the steam that exits the boiler. (The direct-contact condenser is not shown in the figure.)

At nominal operating conditions (i.e., 2.35 gpm of 36% lithium chloride solution, 307,000 Btu/h firing rate), the first-stage boiler removed 153 lb/h of water from the desiccant at a gas-based efficiency of 52.9 % (which corresponds to a COP of 0.95 when a scavenging-air regenerator is added).  The steam leaving the boiler, when condensed, had a solids concentration of less than 10 ppm.  This low level of solids in the condensate places an upper bound of about 6 lb per year for desiccant loss from the regenerator.  This low loss will not create maintenance problems nor will it significantly increase operating expenses.

An energy balance on the boiler showed that heat loss through the insulated jacket was 10%.  This value is much higher than the 2% to 5% that is typical of most boilers and indicates a need to better insulate the unit.  With insulation that brings jacket losses down to 5%, a 1½-effect regenerator that uses this boiler as its high-temperature stage will have a gas-based COP of 1.05.

AILR has developed a manufacturable, beta prototype of the 1½-effect regenerator with funding from the National Renewable Energy Laboratory.  This prototype has a 250,000 Btu/h firing rate and the capacity to remove 250 lb/h or water from a 39% solution of lithium chloride.  The prototype includes both a high temperature and a low temperature interchange heat exchanger. Testing of the prototype was successfully completed in March 2006.  The field operation of this regenerator could begin as early as the summer of 2007.