A venture to manufacture and sell a liquid-desiccant product into the HVAC market can succeed only if customers are confident they are buying a low-maintenance device.  Establishing that confidence, however, is particularly challenging for emerging liquid-desiccant products because of the technology's uneven past history.   While there will always be new O&M requirements for a product that has a liquid-desiccant component, cleaning up messes caused by desiccant leaks and droplets in the supply air must not be one of the them. 

    The industrial liquid-desiccant dehumidifiers rely on droplet filters (i.e., mist eliminators) to capture droplets created as part of the unit's operation.  However, no filter is 100%.  A better solution for a liquid-desiccant HVAC product is to operate under conditions where droplet entrainment doesn't occur, with droplet filters installed only as an added precaution against carryover.

     Although we firmly believe that zero-carryover operation can be achieved, we in no way want to minimize the challenge.  We've had a number of unpleasant surprise in our 35 years of operating liquid desiccant systems, including:

• Capillary pumping of desiccant via narrow grooves and crevices that might be present in the walls of sumps, pans and collection trays; when capillary pumping leads to dipping, droplets are created that can be entrained by the flowing air.

• Flaws within the contact media that lead to dripping within the media even when there is no air flow.

• Contamination of the desiccant that produces foaming, with the concomitant bursting of foam bubbles producing desiccant droplets.

• Dirt accumulation on the desiccant-wetted contact media producing local sites for desiccant to accumulate and drip, again producing droplets.

• Improperly installed pads of contact media created gaps where local air velocities were several times the average velocity

• Air velocities through the desiccant-wetted contact media that are sufficiently high to create shear that rips desiccant droplets off the media's surface.

The last mechanism for droplet formation differs from the others in that it is less a Q/A and O&M problem and more a design problem.

     Although droplet entrainment caused by the shear of the air is less likely at low air velocities, there are practical limits--the face area of the contact-media pad increases as velocity decreases.  So, what is a design face velocity that safely protects against shear-induced droplet entrainment?

     Our design approach starts with the contact media.  We've had success with structured, corrugated contact media made from sheets of non-woven fiberglass, similar to Munters' GLASdek (tm) or Kuul's FirePro (tm).  Although more expensive than their cellulose-based counterparts, the fiberglass is more porous and wicks the desiccant.  We then flood the contact media with desiccant at rates sufficiently low that all liquid flows within the pores of the fiberglass sheets.  As shown in the first video at the right for 3.2 m/s face velocity, when flooded at 0.45 gpm/ft2 (20 lpm/m2)* the media looks damp, but it is difficult to detect the desiccant flow.  With all desiccant flowing within the pores of the media, and none flowing as rivulets on the surface, the desiccant is shielded from entrainment by the air.  Concentration gradients can form in the pore-bound desiccant flow, but we estimate that overall heat and mass transfer coefficients are at least 80% those for water-wetted contact media, where concentration gradients are not a factor.  (It is possible that the random pattern of fibers in the media causes the desiccant to mix as it flows downward.)  

     Droplet entrainment becomes a problem when the shear of higher air velocity first pulls the desiccant out of the pores and onto the surface of the media.  This incipient carryover conditioner is shown in the video for operation at 5.5 m/s face velocity where very agitated pools of desiccant are collecting on the inlet face of the pad of contact media**.  And, as shown in the view of the downstream face of the pad, gross carryover of desiccant occurs at a face velocity of 6.2 m/s.

     For the LDDX described in our LD-DOAS Deep-Dive, pads are flooded at about 0.7 gpm/ft2 and pad face velocities are 1.0 m/s on the absorber and 1.3 m/s on the regenerator. With quiescent pore-flow demonstrated at air velocities 2.5X to 3.0X higher than our design velocities, droplet carryover is not an issue.

     AILR's U.S. Patent 10,655,870 provides a competitive advantage for LDDXs that avoid droplet carryover by adhering to the patent's design limits. The first claim of this patent applies to desiccant-wetted pads of porous contact media that absorb water vapor from air that has been cooled in an upstream heat exchanger when the ratio of desiccant mass flow divided by air mass flow (L/G ratio) is less than 0.147.  

     For the LDDX in our LD-DOAS Deep-Dive, the absorber pad is flooded at about 0.7 gpm/ft2 and its face velocity is 1.0 m/s.  At these conditions, the L/G ratio is 0.052.  For this design to operate outside of the AILR patent, either the flooding rate would have to be increased to 2.0 gpm/ft2, or the air velocity decreased to 0.3 m/s, or a combination of both.  These options all lead to a less competitive product.  Increasing flooding rates will increase the sizes of pumps, sumps and desiccant charge.  But perhaps more importantly, it pushes operation toward a regime where the desiccant flows as rivulets on the surface of the contact media, greatly increasing the possibility of droplet carryover.   And, working outside the patent by significantly reducing face velocity produces an impractical design with unreasonably large face areas.

3.2 m/s face V
Quiescent Flow

(left click on image starts video in separate window)