Is geothermal dead?
![](/app/themes/greenbuildingadvisor/_assets/img/avatar/avatar_50x50.jpg)
I’ve heard people say that geothermal doesn’t make sense any more, because the latest generation of variable-speed mini-split systems has higher efficiency with quite a bit lower costs. Fair enough.
My understanding is that the biggest determinant of heat pump efficiency is the “lift,” the temperature difference between the source and the sink. Ground-source heat pumps traditionally beat air-source on lift, because typically when you’re heating the ground is warmer than the air and when you’re cooling the ground is cooler than the air.
However, the variable-speed systems get an efficiency boost by varying the speed of the compressor so that the hot side only gets as warm as it needs to be for current conditions, minimizing the lift. Most of the time temperatures are milder than design extremes, so most of the time you get an efficiency boost. Conventional heat pumps work at constant speed and modulate by turning on and off; the speed has to be set for extreme conditions, so efficiency suffers in moderate conditions. Currently, variable speed compressors are only available for air-source heat-pumps, and the efficiency gain of variable-speed trumps the efficiency gain of a moderate source temperature.
Here’s my question: is there any reason ground source heat pumps couldn’t use the same technology, varying compressor speed to minimize lift to what is currently needed? And if that were done, wouldn’t geothermal regain its crown as king of efficiency?
GBA Detail Library
A collection of one thousand construction details organized by climate and house part
![](https://www.greenbuildingadvisor.com/app/themes/greenbuildingadvisor/dist/img/modules/detail-library/thumb1.png)
![](https://www.greenbuildingadvisor.com/app/themes/greenbuildingadvisor/dist/img/modules/detail-library/thumb2.png)
Replies
Yeah, it's buried next to the solar thermal......supposedly.
I'm not sure that geothermal has ever relinquished the efficiency crown, because of physics, as you describe. It's just when considering other factors--cost and complexity--it dawns on the wise consumer that this emperor has no clothes.
Waterfurnace has the 7 series which is variable speed compressor tech, also uses variable speed loop pumping.
Costing 3-4x of a mini split system is tough to justify.
I have a 10 yr old GSHP system, back then I don't believe there were good cold climate air source heat pumps available.
Chris- RI USA
I wouldn't say mini splits killed it, although I think cold climate heat pumps in general doesn't help their sales.
I think ground source just loses the cost vs benefit war, and has for a while. I spent tens of thousands extra (guessing 20) on a single 6ton water furnace ground source heat pump for a new house for myself. Lasted 10 years, went through 2 compressors and 2 coils at thousands of dollars out of pocket repair cost - even while "under warranty". On the 3rd compressor failure, this time out of warranty, I replaced it for $15k with a similar ClimateMasters unit. My point being, you have to consider replacement/repair cost as well. And they aren't that popular which means everything you do to them costs a premium. I originally estimated payback at 15 years when the house was new. In hindsight it will never achieve payback. The only saving grace is I got 2/3rds back in rebates and tax credits for the replacement unit.
I thought I was doing the right thing at the time, but as others have said here before, I would have been light years ahead just taking the initial extra money spent on the ground source unit and spending it on improving the building envelope.
The envelope please !! This is where it all starts, all the expensive equipment in the world will not pay back over time like an efficient building envelope. Think of a very high performing building envelope as a legal hidden bank account.
See here, where the lifetime operating cost of variable speed geo @ COP 5.2 is say $4000. Say air source heat pumps @ COP 3.0 would be $7000. The $3,000 difference probably won't cover the additional geo costs (although something like a pre-existing well helps).
On the other hand, the ROI might beat triple pane windows - and people do that to reduce carbon. And add something for the better aesthetics and outdoor noise of geo vs ASHPs. Maybe less demand charges too.
Jon has a good perspective. It might not win the financial argument currently. But lots of things don’t. People spend all kinds of money on cosmetic and luxury things. Some will get more satisfaction from a more efficient heat system and others from a fancy stone facade.
I do think the the equipment itself has some catchup to do with variable speed compressors and such. But I suppose they need to get the price down more than they need higher tech equipment.
You have a point, but those extra savings on $ could be put towards more insulation or something else that might result in more overall energy savings than you’d get using a ground source heat pump. It’s important to think of the house as a complete system, and try to allocate your money to get the maximum overall benefit.
I do wish ground source heat pumps were cheaper to install as I’ve always thought they are a really nifty way to heat and cool a home.
Bill
Thanks for the link, that's a useful table. One of the things that jumps out at me is that the COP of 5.2 is with an open loop, that same machine with a closed loop is more like 4.3. Everything I'm reading is that open loop systems are just a lot more trouble, closed loop is preferable. Now what I really want is a hydronic system, but the best COP I see for that is 3.5.
With a closed system, all you really need to do is mix in enough glycol to prevent freezing, and some corrosion inhibitors. Periodic (commercially it’s done annually) checks of the corrosion inhibitors is about all you need to do maintenance wise, and add in anything that may be missing. The corrosion inhibitors get used up over time.
With an open system, you need to provide make up water to replace water lost to evaporation, corrosion inhibitors are a problem since the system gets oxygen in it from the “open” part that is exposed to the atmosphere, and you get all kinds of detritus in the system like insect parts, tree gunk, and pretty much everything else. There is a lot more maintenance with an open system.
Bill
Exactly. But the same heat exchanger rates a full point lower in COP for closed loop vs. open loop. If the trade-off with an air-source is that you're trading higher efficiency for higher initial cost, if you then have to decide whether to give up some of that efficiency for the sake of reliability it kind of ruins the deal.
" One of the things that jumps out at me is that the COP of 5.2 is with an open loop, that same machine with a closed loop is more like 4.3. Everything I'm reading is that open loop systems are just a lot more trouble, closed loop is preferable. Now what I really want is a hydronic system, but the best COP I see for that is 3.5."
Do you think the difference in the COP of same equipment connected when connected to open vs. closed loop has to do with the open loop assumes it is feed by an artesian well so no electricity is used for pumping the water?
From what I can tell no two loops are the same, this one this one is 3 vertical wells 200 feet deep the next is 2 400 foot wells the next one is horizontal with all different lengths
Can you really say every loop will use the same amount of electric to move the water?
If you do not know how much electricity you will use to move the water how can you say what the COP is or will be?
Is the COP different in November when the loop temp is 55 or 60° than in February when the loop temp 30 or 32°?
I am sure the industry and government has a 3000 page manual on how to measure COP. I am also sure 99.9% people buying equipment do not understand the manual or what the COP number means.
Walta
Thanks Walta.
Reading a spec sheet, I noticed the open loop spec assumes incoming water temperatures of 59F for cooling and 50F for heating. The closed loop spec assumes 77F and 32F. That's a big difference.
I don't know if that's industry standard for rating.
Walta....LOL! Those nutty numbers....
"Walta....LOL! Those nutty numbers...."
Yes the 3000 page manual was a joke.
But the water temp, wells and well depth numbers seem plausible after the time I spent reading on the Geo Exchange forum.
If my build would have happened at a time when the tax credits were available I would likely have one.
https://www.geoexchange.org/forum/
Walta
Bringing the conversation back around, I think Jon has put a good point on things here. If the lifetime operating cost of an air source heat pump is $7,000, the most that could ever possibly be saved is -- you guessed it -- $7,000. The more that number comes down, the harder it is to justify higher efficiency at higher cost. And in the world of geothermal that's not a lot of money.
I would think that a big chunk of the extra cost of a geothermal system is the well or trenches required, and that's not going to come down in price.
OK, two questions:
1. The secret sauce of the mini-splits seems to be in two parts. First is the compressor with a variable speed. Second is at the head. Ground source heat pumps are now coming out variable-speed compressors. Could they be paired with mini-split style heads?
2. Could a mini-split style head be run with water as the medium rather than refrigerant? A water-to-water heat pump is a simple box, water in and water out, and minimal refrigerant. You don't need an HVAC tech to install it, a plumber or even a competent handyman can run water pipes. The box itself is a closed package, no charging and less maintenance. This would allow ground-source to get some cost advantage back over air-source mini-splits that have to run line sets to each head. Is this practical?
I hope someone answers yes. What you are describing sounds like a neat system if it would work.
I can’t really answer #1 since I don’t know enough about the specifics of these systems. My guess is “yes”, if the control system was compatible.
I CAN answer #2, and that is unfortunately a NO. The reason is because the phase change (liquid to gas or vice versa) that you get with refrigerant in the heat exchanger coils allows for VASTLY more thermal energy transfer than you can get with an all-liquid system using water. It’s somewhat similar to the difference between hot water and steam heating systems. This is just physics showing up to ruin our fun.
If you were to run with a water to air heat exchanger, you’d have a passive system which WOULD provide some cooling or heating ability based on the thermal differential between the water and the air temperatures, but it would be much, much less useable BTU capacity in the units. You’d probably lose all dehumidification capability in Cooling mode too. The chilled water systems I design with at work run around 45 degree F supply water temperatures and pretty high flow rates with large pipes to handle the volume of water required.
If you are using 50ish degree supply water from an in-ground pipe, you’d never see frost like you can see on refrigerant lines, for example, and you’d never be able to heat your space above a bit less than 50 degrees. The other downside is that as your room temperature approaches the water temperature, less energy gets transferred due to the lower temperature differential. You also have no ability to cool to lower than the temperature of the water (probably no big deal, who wants a 50 degree F house?), but you also can’t heat your house any warmer than that water temperature.
A passive system doesn’t have the ability to force thermal energy against the gradient (to heat a house with heat scavenged from cold, outdoor winter air, for example) without the phase change you get with refrigerant and a compressor in an active system such as is used in a minisplit setup.
Sorry to be the bearer of bad news guys.
Bill
Just to clarify, I'm not talking about a passive system. In a commercial setting what I'm talking about would be called a "water-cooled chiller," with the difference that the cooling water would come from a geothermal well as opposed to a cooling tower. The chiller is a heat pump with refrigerant in it, but instead of outputting hot or cold air it outputs hot and cold water from the two sides.
Let's say the chilled water is leaving the chiller at 30F (it has antifreeze) and returning at 55F. The coil would have an average temperature of 42.5F which should give ample dehumidification. With a 25F temperature delta a one-ton system would require 480 pounds of water per hour, or 60 gallons, or one gallon a minute. I think that's in the territory of 3/8" PEX. On the ground loop you'd have the same 1 gpm with the same 25F delta, so you could be taking water in at 55F and dumping it at 80F.
What I don't know is where the special sauce lies with the mini-split, is it in the cooling head? Chiller-based cooling systems are widely used in commercial buildings but the air handlers tend to be rudimentary. Or is it in the evaporator? That would be even better, because in a water-water system the compressor, condenser and evaporator are all in the same box.
I work with chillers like this at work. They are the most efficient way to handle very large cooling needs like the datacenter facilities I design. They usually dump the heat in a cooling tower, where evaporation helps with the cooling (think of a cooling tower as a watefall with a big fan blowing through it), or large cooling ponds which are exactly what they sound like. I’ve never seen a system using a geothermal loop, and I’d guess it’s due to the size of the geothermal loop that would be needed to handle the load. The smallest cooling plant I’ve worked with is a 350 ton system. Most are much larger, often 1,000 tons or more.
The air handlers are basically radiators in reverse, there isn’t much to them. Some have reheat coils that heat the air before entering the cooling air to allow for more dehumidification, but that’s about it for fancy stuff. The newer air handlers use what are know as “EC” fans (I think it stands for “electronically commutatated”), which are variable speed blowers, basically.
Normally the air handlers modulate cooling water flow through the coil using either a three way “bypass” valve (water either goes to the coil then the return, or gets diverted directly to the return so the flow rate to the system remains constant), or just a regular valve that means the flow rate to the system is variable. Usually the airflow remains constant, but in newer systems with EC fans the airflow varies. This is a bit like a minisplit, and allows full control of delivered BTUs or cooling.
A system with a chiller should work as you describe, but I’m not aware of any such systems. Water cooled chillers do not like to cycle though, they are setup to run continuously and modulate capacity by either changing the pitch of the compressor blades or changing the speed of the compressor motor (in the case of variable speed drives). You generally try to design these systems so that the chiller runs constantly and never cycles. That may be part of the reason such a system hasn’t been adapted to smaller residential systems.
BTW, cooling water out of a chiller is never below freezing in normal building cooling systems. 45 degrees F is a common design temperature, I’ve see 42 and 40 but never lower. Typical thermal differential across an air handler is usually 10 degrees, but may be 20 or 30 degrees in some system designs. Higher differential means more BTUs, but usually different sensible/latent combinations. In datacenters we only care about sensible cooling capacity.
Bill
I've done controls work on commercial Mitsubishi water source VRF system coupled to ground loops here in Canada. They exist and work fine.
The reason why geo isn't used on data centers or other process style loads, is often because it is so cooling dominated. Typical geothermal ground loops depend on being reasonably balanced between heating and cooling, over the course of a year, or heating biased. If it is too strongly biased towards cooling the loop tends to heat up overtime and dry out the surrounding soil which reduces the thermal conductivity and kinda starts a slow death spiral of reduced capacity. In Canada, residential loops do fine because they are heating dominated which tends to hold the moisture in the vicinity of the loop, and there is enough of a seasonal rest to be ready for the next winter. But many commercial loops have some sort of evaporative fluid cooler to keep the loop at long term sustainable temperatures.
And the economics of a multi thousand ton ground loop would be not palatable.
This sounds similar to what https://www.chiltrix.com/ are doing with air-source, but with a geothermal hookup instead.
Thanks Bill, good to hear first-hand knowledge, you know more about this than I do. The EnergyStar website (https://www.energystar.gov/products/energy_star_most_efficient_2019/geothermal_heat_pumps ) lists four manufacturers of of water-water heat pumps, but three of them seem to be rebadges of the same unit, so it looks like there are really just two being made, in a few different sizes.
There's the GeoComfort/Hydron/Tetco (see https://geocomfort.com/residential-products/category/guide-wv-variable-speed)
Which looks to be available in 2 ton and 3.5 tons (cooling, slightly more heating), maximum COP of 3.9.
Modine makes both heating-only and heating-cooling systems. These were previously sold under the Geofinity name, which Modine acquired. See http://igate.northernplumbing.com/specsheets/modine/ghr.pdf ( couldn't find anything on the Modine website). Seven models from 3 to 11 tons.
No matter how sophisticated the compressors & controls, there's no getting around the installed cost difference between a ground loop for a GSHP and the cost of an air coil on the outdoor unit.
The fact that the ground heat exchange is (almost) always a custom design, there is also some design and implementation risk that can bring the as-used COP to it's knees.
Yes, in an ideal installation in a cold climate the COP at peak load conditions the GSHP can be 2-3x more efficient than a modulating ASHP, but the difference under average conditions may not be all that great (in a poor implementation it can even be less efficient), losing the cost rationale for that enhanced efficiency.
For many MANY real-world implementations it's more cost & comfort rational to spend the installed cost difference on improving the thermal efficiency of the house, rather than on the efficiency of the mechanical systems.
In New York's Hudson Valley the Google spinoff Dandelion Energy (https://dandelionenergy.com/ ) is making an attempt at lowering the costs by keeping it to a few cookie-cutter designs using lower cost drilling techniques, and using economies of scale, such as pre-drilling a whole city block's worth of wells (even in front of homes that have not contracted for the GSHP retrofit), which seems to be working OK in neighborhoods that don't have access to the gas grid. It's still more expensive than ASHP solutions of comparable capacity, but with the remote monitoring to detect/correct maintenance and installation issues and the access to cheap capital allowing them to provide cheap financing it's not necessarily a worse deal than an ASHP solution. They've taken a cue from the home-solar industry with various options of doing it with no or low money down.
When viewing some of their installation videos and customer testimonials I cringe a bit when I see how much they COULD have spent fixing the thermal performance of the house though rather than adding another ton or so of GSHP to the retrofit. Spending 3-5 grand on air sealing and insulation in some of the those houses probably would have saved more than that on the upfront installation cost, no matter how creatively financed.
I think the only way to get small geothermal systems comparable to air source systems in terms of cost is to share some of the infrastructure, specifically the excavation part. I just don’t know if that’s possible or practical.
Dana, I’ve often thought to use some of the cooling tower water on the “hot” side to circulate through some PEX loops in parking lots on facilities I design. The idea would be to use some of the waste heat to thaw snow and ice, thus avoiding the need for plowing and salting the parking lot. Do you think it a residential setting it might make sense to try putting some PEX loops into an asphalt driveway and using that as the geothermal loop? While not buried, the asphalt driveway might act like a large solar collector during the day, and would be a pretty big thermal mass to hold some of that heat into the night. If the driveway was going in anyway, the PEX could be installed between layers of asphalt and save some of the install costs. I just don’t know if there would be enough of a benefit to make it worthwhile.
Bill
The asphalt driveway as solar collector works pretty well for pool heating in the summer (even without a heat pump) but probably isn't going to cut it for space heating in locations that see much snowfall.
But as a heat dump for chillers on server farms it should work pretty well during the cool dark winter, but you'd still, need something else for the shoulder seasons. Whether it's "worth it" requires both a technical and financial analysis. Most places don't need snow melting at a high duty cycle, but at the heat of fusion of water it's a great way to dump a lot of heat all at once.
The datacenter parking lots aren’t big enough to ever handle the entire load. One of the battles we always get into with the cities is the number of parking spots isn’t big enough for the square footage of the building. We have to explain that there is only a very small staff in the building. The cooling loop in the Parking lot would only be used in winter months, and would still only have some fraction of the total system diverted to it. The cooling towers would still be active.
We typically run evaporative cooking towers on the larger facilities and drycoolers on medium ones. Small facilities get regular Freon systems, but some of those now have economizer setups where the Freon can circulate a bit without the compressor running if it’s cold enough out.
I’m always looking for a way to squeeze out another percent of efficiency somewhere. All the easy stuff has already been done, so now we think creatively. We’ve actually seen Mutli-thousand dollar monthly savings by building sliding air baffles to use as doors at the ends of rows of equipment.
Bill
I am sitting in an under construction (mostly occupied, it's going in phases) higher ed building. This uses 'chilled beams' which seem to be some sort of air handler that is sized to fit a standard ceiling tile grid. These are for heating and cooling. Some of the open spaces have Runtal radiators along the windows. The interesting part is that *some* of the heating/cooling is coming from two wells that go through the silt (Missoula flood, x1 or x6 depending on how you talk to) and down into Spokane/Rathdrum aquifer. The control equipment side in the building is interesting, a lot of gear. The water is pumped up to grade, exchanges heat, then is pumped back down the other well. The two well heads are 20 feet apart. I guess they can move the whole building temp 2 degrees either way with the aquifer water. Not sure what the efficiency is or if it was even worth it.
The aquifer is 10trillion gallons, so dumping some slightly warmer/colder water back down probably isn't a big deal.
Controls have gotten cheap enough that it’s economical to implement complex control systems to allow for more system optimization including adaptive optimizations.
The hard part is trying to get the energy to do the actual work of heating or cooling. Geothermal’s main advantage is that the typical 50ish degree ground temperature keeps the “outdoor” side of a heat pump are near optimal operating temperatures over an entire year so more efficient heating especially, but also cooling. The problem is the cost of putting the “geo” part in the ground.
Bill
You might find this interesting. Citrus in Neb during the winter. Passive (somewhat).
https://youtu.be/ZD_3_gsgsnk
I just stumbled upon a Mitsubishi ground-source heat pump with variable refrigerant flow (VRF):
https://www.mitsubishipro.com/pdfs/6-geothermal.pdf
They don't seem to be available in the US though.
Commercially Mitsubishi offers water source units that have a minimum inlet water temperature of 23f. These are in the US mylinkdrive page.
http://meus1.mylinkdrive.com/files/PQHY-P72TLMU-A1_208-230V_Submittal-en.pdf