I’d like to talk about dehumidifiers; what they are, how they work, design failures and improvements I’d make.
The dehumidifiers I write of are a specific kind, but probably the most common. They are the small white plastic towers people role around their basement, seeking out the damp spots, that are typically powered by 110 VAC.
Sure, there are other ways to dehumidify an area; desiccants come to mind, but these little units are usually relatively cheap and are effective for 10 to 20 years. This type of dehumidifier is pretty common in North American households. I’m not aware of the technological state of these units in Europe, but I would imagine they are quite a bit more efficient; seems most things are. Most people I know that use these to dry out a damp basement, closet, coat room or as I recently learned, many people dry their clothes this way.
These machines operate by a closed loop vapor compression system that cools the water vapor in the air below the dew point causing it to condense out there by lowering the specific humidity in the room. Internally, at least in the machines I’ve seen, they have the same basic components found in a typical household refrigerator or window air conditioner. It consists of a compressor, discharge line, condenser, capillary tube, evaporator and suction tube. The refrigerant in the machines I’ve seen is R-134a, but R-22 used to be common and probably R-12 before that.
The general idea when it comes to controlling humidity in a space with a vapor compression dehumidifier (after options like proper ventilation and appropriate insulation have been exhausted), is to circulate the humid air across a cool coil absorbing heat from the water vapor to the point where it condenses on the coil and drips into a bucket or runs down the drain. This is not an easy task because of the large amount of heat that must be removed to cause this condensation. Water has one of the largest vaporization latent heats of any common substance at 2260 kJ/kg. There is more heat released by the condensation of a liter of water than than a 600 watt heating element running for an hour! Dehumidifiers condense out several liters of water a day. All that heat must be discharged somewhere, like a refrigerator or air conditioner. A window mounted air conditioner is very similar to a dehumidifier except that it discharges the heat of condensation outdoors by passing ambient outdoor air across the condenser coil. A dehumidifier on the other hand is like a window mounted air conditioner where the cooled, dry air after passing through the evaporator, also passes the condenser; picking up the heat of condensation.
There really is not much to them. Because the heat of condensation is returned to the cool air rather than what might be much warmer air outside or even just the ambient air in the room, the temperature that must be maintained in the condenser can be much lower than an air conditioner and so then the pressure too. A lower head pressure is great for the system performance and means a smaller system to do the same job or a greater capacity for a given size. Because the heat of condensation is released in the room, plus the heat of compression, the temperature in the room generally rises. In some cases this is desirable while in others, not so much. At the same time the air is also now drier. Some people like to say that warm air “holds more water” than cold air; this is sort of true. In any liquid/vapor environment there are always molecules jumping from liquid state to vapor and back again. The amount in each state depends on a number of things, but most important are temperature and pressure. At higher temperatures more molecules of water will “jump” into vapor state than jump back into liquid so the air “holds” more. At lower pressures the same thing happens, but that isn’t too important in this discussion except inside the vapor compression loop itself; more so in meteorological study and weather prediction. The engineering discipline concerned with these vapor/liquid relationships is called Psychrometrics. There is a common chart by the same name often used to determine appropriate human comfort levels relating to the HVAC industry.
Now here’s where it gets dirty. Aside from the toxic, global warming, proprietary refrigerants and the inefficient, un-servicable compressors, there are some design considerations I would like to address. These small dehumidifiers are mass produced to be as cheap as they can be while providing a lengthy enough service life to merit a replacement. I will admit that for the original material cost as well as economic, these little machines provide a better bang for your buck than something similar 40 years ago. Capitalist competition, in trying to produce a nearly equivalent product for a cheaper price, has engineered the cheapest inefficient machines that will perform satisfactorily for a predetermined calculated lifespan. But I digress; the same set of principles applies to most consumer goods.
One of my biggest pet peeves, after improving the compressor, using more sustainable materials and reducing or eliminating superheat coming back to the compressor, is the practice of overcooling condensate. The refrigerant control in a dehumidifier is almost always a capillary tube. It’s just a very small bore copper tube maybe a meter long that high pressure liquid refrigerant is forced through to the evaporator on the other end. The designers of these machines calculate the appropriate bore and length of tubing to get the needed pressure drop within the design pressures of the whole system so that the refrigerant, upon entering the evaporator, will boil at a sufficiently low temperature to absorb heat from the humid air passing across the coils. Below a certain temperature called the “dew point”, the amount of water molecules vaporizing is less than the amount condensing and so the condensation forms and the water beads up and drips into the pan. This is all well and good, but some extra work has been done which is cooling that condensate dripping into the bucket. You might not think it is much, but the heat capacity of liquid water is pretty high and I feel this is a considerable blunder. Worse yet, these machines are designed for rather specific operating temperatures usually far lower than it needs to be ensuring humidity is condensed from the local environment. This leads to even cooler condensate, lower suction pressures, higher head pressure and an overall loss in efficiency. Granted, these machines probably have a C.O.P. of better than 2.0, potentially it could be greater than it is. I see overall efficiency as the actual C.O.P. relative to the ideal system. Ideally, we want to operate the evaporator at a temperature just low enough to effectively condense water from the air, but not so low that we overcool said water. To achieve this I think that most engineers would complicate the apparatus with a form of a Thermal Expansion Valve or a lot of electronics and an Electronic Expansion Valve. Both could be made to work, but are rarely seen in household appliances because they add expense to the initial cost of the unit. One thing that could be done to counter the losses to the thermodynamic refrigeration cycle is to use the overcooled water in a similar manner to using the chilled, drier air from the evaporator to carry heat away from the condenser. Achieving significantly lower condenser pressures by cooling it in the water condensate directly isn’t a practical option considering the quantity of heat to be released and the small amount of condensate to absorb it would be heated rapidly to a high temperature. There are other heat sinks for this condenser heat in a common household such as a preheater for a domestic water heater or even heating the greywater effluent exiting a home.
But as far as utilizing the thermal potential of the over cooled water condensate, I propose a very simple change in plumbing of the end of the condenser coil that would subcool the pressurized refrigerant close to the evaporator temperatures. This would significantly reduce the amount of flash gas in the capillary tube and increase the overall refrigerating effect of the cycle.
Another idea I hope to draw out and share sometime would use the condensate itself to act as the condensing surface by which further condensation takes place. This might be done by refrigerating an insulated container of water well below the dew point of the ambient environment and pumping this to the top of a two meter tall, 4o centimeter wide pipe where it is discharged and slightly atomized falling through a column of air blown through from the room. The idea is that at some point in the air column, the chilled water will condense out more water from the air which raises the temperature of the water droplet, it then falls back to the water storage to have the heat removed by the submerged evaporator coil. That’s not the end of it though. If the original conditions of the water (temperature, volume, flow rate and diffusion) were well designed, then the falling water droplets could continue to condense out water from the air stream, slightly rising in temperature as it fell. The rising air would have a lower dew point the further up the column as it meets colder water droplets cold enough to continue the condensation. The ambient air entering the bottom of the column with the highest dewpoint meets water droplets warmest, but still cold enough to condense water out. The overall picture is a temperature gradient forming with the dewpoint of the air falling as the air rises and the temperature of the water rising as the water falls. Instead of circulating the air in a room through a conventional dehumidifier many, many times; lowering the dew point slightly with each circulation, this would instead dry the air out in fewer, maybe even one, pass. I have no idea if this thing would work, but it’s a fun idea I might toy around with sometime.
Above, is a dehumidifier that ran in my basement for about a year. Someone gave it to me because it didn’t work. It wasn’t very old so I looked it over and found a puncture in a copper line where it must have been bumped up against something and was smashed. I repaired it and replaced the capillary tube with a brass flow control valve seen on the left. I had to put a different fan on it and make a new shroud since the original fan/shroud was part of the outer plastic shell. I charged the machine with propane sourced from a 20 pound propane cylinder. It worked great although I would have to periodically adjust the refrigerant control to get the most output per kilowatt-hour which I monitored over the course of it’s use. It was a good proof of concept, but wasn’t easy to collect temperature data from it to build a model of it’s operation.
I think this is the same dehumidifier before I put it to service in my basement. Here, I bypassed the original condenser and rigged up a submerged condenser in a larger Coleman cooler full of a known quantity of water. I monitored and logged certain system temperatures as well as the temperature rise of the water in the cooler over a predetermined time period. I used this information along with the power consumption and mass of condensate produced to develop a crude model to evaluate system characteristics and the affect of various changes.
I’d like to do more serious work in developing a better dehumidifier someday, but my primary interest is domestic refrigerators.
[…] Dehumidification may be done in a similar fashion, but a small circulating brine loop and fan, could be a better option. […]
So this has been up for awhile obviously but I have a couple questions if you happen to check these comments.
1) ” I propose a very simple change in plumbing of the end of the condenser coil that would subcool the pressurized refrigerant close to the evaporator temperatures” What was the plumbing change? Are your routing the cold condensate back to just after the condenser? Basically dripping cold water on the refrigerant carrier pipe after’s been cooled by the condenser to drop the refrigerant temp even more without also re-vaoprizing the condensate water?
2) Did you ever build the dehumidifier cooling tower version?