Compressor Based Liquid Desiccant Air Conditioner

LDDXConcept2.gifThe Liquid-Desiccant Direct-Expansion air conditioner (LDDX) is an efficient, high latent cooling system that can adjust its Sensible Heat Ratio (SHR: the fraction of total cooling that is sensible, the balance being latent) between 0.35 to 0.75.  When operating at ARI A test conditions, the LDDX’s SHR can be controlled with minimal loss in its rated total cooling capacity while its efficiency stays close to its projected value of 11.4 EER.

Peterson's and Howell’s U.S. Patent No. 4,941,324 and their ASME Paper 86-WA/Sol. 9 (December 1986) are probably the first description of an LDDX.  Much of the early testing by Peterson and Howell used a glycol as the liquid desiccant.  This choice of a non-corrosive desiccant allowed the LDDX to use conventional aluminum-fin heat exchangers.  Unfortunately, the intrinsic vapor pressure of glycol prevents its use in HVAC applications. Following their initial report on the glycol LDDX, Peterson and Howell appear to have stopped work on the concept.

To our knowledge, AILR’s LDDX with wicking-fin technology is the first embodiment of a compressor-based liquid-desiccant air conditioner in which a solution of lithium chloride floods the surfaces of the air conditioner’s evaporator and condenser providing direct contact between the desiccant and the air flowing through these two coils.  This implementation of an LDDX is shown in the neighboring figure.   As shown in this figure, the LDDX’s refrigerant circuit functions identically to that of a conventional DX air conditioner: (1) a compressor supplies high pressure refrigerant vapor to a condenser, (2) the air-cooled condenser rejects heat as the refrigerant condenses, (3) the liquid refrigerant partially flashes to vapor as its pressure drops across a thermostatic expansion valve (TXV), (4) the two-phase refrigerant mixture enters the evaporator where it fully converts to vapor as it absorbs heat from the air flowing through the evaporator, and (5) the compressor maintains the evaporator at low pressure by pumping refrigerant vapor from it, thus completing the cycle.

Integrated into this refrigerant circuit is a liquid-desiccant circuit that supplies strong desiccant to the top of a wicking-fin evaporator and weak desiccant to the top of a wicking-fin condenser.  The strong desiccant absorbs water vapor from the process air flowing through the evaporator.  The weak desiccant flowing off the evaporator is warmed in the interchange heat exchanger before it is delivered to the top of the condenser.  The heat rejected in the condenser further warms the weak desiccant which then desorbs water to the cooling air that flows through the condenser.  The warm, strong desiccant that flows off the condenser is cooled in the interchange heat exchanger before it is supplied to the evaporator.

LDDXDesiccantRecirc.gifThe essential characteristic of the LDDX is its ability to supply cool, unsaturated air without reheat.  Whereas a conventional DX air conditioner without reheat might have a sensible heat ratio (SHR) of 0.75 (i.e., 75% of the cooling it provides is sensible temperature reduction and 25% is latent dehumidification), the LDDX can have an SHR as low as 0.35.

The LDDX can also adjust its SHR over a fairly wide range so that varying latent and sensible loads within the building can be independently served (i.e., temperature and humidity can be independently controlled).  This adjustment is achieved by varying the amount of liquid desiccant that is recirculated over the evaporator.  In the neighboring figure, a diverting valve splits the weak desiccant that leaves the evaporator into a stream that flows to the inlet of the pump that supplies desiccant to the condenser and a stream that is returned to the evaporator.  As the fraction of desiccant returned to the top of the evaporator increases, the concentration of the desiccant becomes weaker and the amount of moisture absorbed from the process air decreases.  The overall effect of increasing the recirculation rate to the evaporator is to increase the LDDX’s SHR.

COPandSHRvsSplitLDDX.gifThe LDDX with a controllable diverting valve will not require an interchange heat exchanger if its operation is limited to relatively high amounts of recirculation (i.e., 60% or more of the weak desiccant collected at the bottom of the evaporator is returned to the top of the evaporator).  This is illustrated in the following two graphs both of which are based on AILR's computer model of an LDDX.  In these graphs the parameter “split” is the fraction of the weak desiccant collected at the bottom of the evaporator that is returned to the top of the evaporator.  The blue line in the graphs shows operation without an IHX and the red points, operation with an 80% effective IHX.

The net COP shown in the first graph is the ratio of the total air cooling provided by the LDDX divided by the compressor power when both parameters have the same units.  Fan and pump powers are not included.  As shown in this first graph, the significant 21% drop in COP that occurs when the IHX is omitted and the LDDX operates with no recirculation decreases to a 2% drop when the recirculation split is 0.8.

The second graph shows that the IHX has minimal impact on the Sensible Heat Ratio of the LDDX.  Furthermore, the SHR can be adjusted between 0.33 and 0.52 while keeping the LDDX’s efficiency within 5% of its peak value by controlling the desiccant recirculation between 0.6 and 0.95.

An interesting operating point for the LDDX occurs when the recirculation of desiccant over the evaporator is 100% (i.e. split equals 1.0).  Although it is unlikely that one would try to operate the LDDX at this condition, it does represent a physically attainable state in which infinitely weak desiccant (i.e., water) flows over the evaporator and concentrated desiccant flows off the condenser at the same concentration at which it flows on.  While there is no net exchange of water at the condenser, water is produced at the evaporator (as would happen at a conventional evaporator that was operating below the dewpoint of the processed air).  Operating at 100% recirculation does increase the concentration of the desiccant flowing off the condenser.  Depending on ambient humidity, crystallization of salt could occur.  But it is interesting to note that at 100% recirculation and 95/75 F DB/WB ambient conditions, the LDDX computer model predicted a strong LiCl concentration of 40%, which is several points below salt saturation. Under the conditions modeled, the LDDX with 100% recirculation has a very high SHR (i.e., 0.88) and relatively low efficiency (i.e., a compressor-based COP that is 22% below the LDDX’s peak value.)

Although a wicking-fin LDDX is not the only way to implement a compressor-based liquid-desiccant air conditioner, it probably is the most efficient.  Alternative configurations that have either been introduced to the market or are in development include designs where:

    • strong desiccant is cooled in a desiccant-to-refrigerant evaporator,
    • weak desiccant is heated in a desiccant-to-refrigerant condenser,
    • the cooled desiccant is delivered to a porous bed of contact media where it dries and cools the process air,  and
    • the warmed desiccant is delivered to a second porous bed of contact media where it rejects water and heat to a stream of cooling air,

and designs where:

    • a chiller delivers cool water to an internally cooled liquid-desiccant conditioner,
    • heat rejected at the chiller’s condenser maintains a warm water loop that recirculates through an internally heated liquid-desiccant regenerator, and
    • weak and strong desiccant are exchanged between the conditioner and regenerator through an interchange heat exchanger.

The two preceding compressor-based liquid-desiccant air conditioners will have relatively high pumping powers—the first for pumping large volumes of desiccant and the second for pumping large volumes of water.  They also will be penalized for the additional temperature drops that are introduced by their fluid circulating loops.  Although manufacturer's performance data is not available for either of these two air conditioners in sufficient detail to make a meaningful comparison, the wicking-fin LDDX described here should be the more efficient high latent option.

EERvsSHR.gifThe conventional, high latent alternative to the LDDX is a DX air conditioner that has a reheat coil immediately downstream of its evaporator.  At least one HVAC manufacturer has implemented this reheat option as a dual refrigerant circuit with staged compressor operation.  A first stage compressor is part of a refrigerant circuit that can be switched between a configuration where all heat is rejected outdoors (i.e., no reheat operation) and a configuration where the hot refrigerant gas from the compressor partially condenses in a coil located downstream from the evaporator before fully condensing in an outdoor condenser (i.e., reheat operation).  The second stage compressor is part of a conventional refrigerant circuit with an indoor evaporator and outdoor condenser. At ARI A test conditions with both compressors operating a 7.5 ton model of this high-latent air conditioner operating without reheat would have a gross cooling capacity of 93,000 Btu/h, an SHR of 0.73, and an EER based on compressor power of 15.8.  Switching to reheat reduces sensible cooling by 25,000 Btu/h while leaving latent cooling almost unchanged.  With gross cooling capacity reduced to 68,000 Btu/h but latent cooling unchanged at 25,200 Btu/h, the air conditioner's SHR drops to 0.63.  Total compressor power is slightly less in reheat mode since the first stage condenser is now larger, but the loss of cooling capacity still drops the EER based on compressor power to 12.1.  When sensible loads are very low but dehumidification is needed, the second stage compressor can be turned off and the first stage circuit operated in reheat mode.  Under these conditions, the SHR drops to 0.19 and the compressor-based EER drops to 8.3.

With the caveat that many factors influence a cooling system's efficiency so an absolute comparison of EERs between the LDDX and a conventional DX air conditioner with reheat can be misleading, the trends in EER as each high-latent air conditioner is modulated to meet different SHR is shown in the neighboring graph.  The three operating states previously described for the conventional "reheat" air condition are shown in this figure.  Dotted lines connect these states indicated that SHRs that fell between the three discrete values could be met by cycling the air conditioner among operating states.

The LDDX's EER peaks at an SHR of about 0.37.  The LDDX's SHR can be increased by increasing the amount of desiccant recirculated to the evaporator, but this recirculation leads to a modest drop in evaporator temperature, which in turn reduces both the cooling capacity and the EER of the LDDX.  The LDDX's SHR can be further reduced below 0.37, but the relatively high exchange of desiccant between the condenser and the evaporator will lead to a sharper drop in EER unless an interchange heat exchanger is included.

Although it is premature to compare capital costs for the LDDX and a conventional DX air conditioner with reheat, it is noted the LDDX may be the less costly alternative when costs are expressed in terms of dollars-per-latent-ton of cooling.  The 0.73 SHR of the conventional air conditioner at ARI rating conditions leads to a capital cost expressed in terms of latent cooling capacity that is 3.8 times higher than its nominal cost expressed in terms of total capacity.  By comparison, the LDDX, with a SHR of 0.37 would have a capital cost  expressed in terms of latent cooling capacity that is only 1.6 times higher than its nominal cost.

TwoStageLDDXDOAS.gifIn applications where a conventional DX air conditioner can maintain comfortable indoor conditions without ever resorting to overcooling-reheat, the LDDX does not offer a performance advantage—in fact, it might be disadvantageous to use an LDDX due to pumping power for the desiccant and the heat load on the evaporator imposed by the warm, strong desiccant. Thus, in applications where high dewpoint air is processed (e.g., Dedicated Outdoor Air Systems operating in humid locales), the most efficient cooling system will be one that divides the cooling process into two stages: a first-stage conventional DX air conditioner that delivers saturated, partly cooled air to a second-stage LDDX.  This configuration is shown in the neighboring figure.  As shown here, the latent cooling capacity of the second-stage LDDX can be increased by placing the condensers for the two stages in series on the air side so that the air entering the LDDX’s condenser is warmer.

In May 2014, AILR completed laboratory tests of a 3-ton LDDX prototype as part of an award under the Department of Defense's Environmental Security Technology Certification Program (ESTCP Award).  This prototype will operate for the second half of the 2014 summer at an Army base in New Jersey. Excerpts from the test report's Executive Summary are as follows:

The laboratory tests confirmed that an LDDX that used wicking-fin technology for its evaporator and condenser can be a highly efficient air conditioner with an EER of 12 that provides almost two-thirds of its cooling as latent cooling (i.e., dehumidification) while supplying air with a dewpoint approaching 46oF. Furthermore, the LDDX can be controlled so that the fraction of cooling that is latent varies so that temperature and humidity within a building can be independently controlled.

The LDDX prototype that was tested in this study was the first implementation of any cooling system using wicking-fin technology.  The uncertainties inherent with the early operation of any new technology produced a prototype that did not effectively allocate heat transfer area between its evaporator and condenser.  In effect, the evaporator was undersized for a three-ton cooling system and the condenser was oversized.  With this “compromised” design, the LDDX prototype achieved an EER of 9.3 and a Latent Heat Ratio (LHR—which is the inverse of Sensible Heat Ratio) of 0.68 and supplied air with a dewpoint of 47oF (approximately 5oF below the suction temperature of the evaporator).

The test of the first LDDX prototype has provided the detailed heat and mass transfer data necessary to more effectively design future units.  The previous projection of an LDDX with an EER of 12 reflects the potential of the technology using experimentally proven design data.

Although laboratory tests have been relatively short (tens of hours), the LDDX prototype has operated with no entrainment of desiccant droplets as the air flows through either the wicking-fin condenser or wicking-fin evaporator.  “Zero carryover” operation has been confirmed both visually (one wall of the condenser exit plenum has a window through which the condenser’s exit face can be viewed during operation) and with a particle measuring instrument located in the supply air stream from the evaporator.  Measurements of airborne particles in the range of 0.5 micron to 10 micron showed that LDDX’s wicking fin evaporator is a moderately good particle filter for larger particles (89.6% capture at 10 microns), but it is a poor filter for small particles (5.4% capture at 0.5 microns). At all sizes, the density of particles in the air leaving the evaporator was lower than in the entering air, which is a strong indication that desiccant droplets are not being entrained by the process air.