GBA Logo horizontal Facebook LinkedIn Email Pinterest Twitter X Instagram YouTube Icon Navigation Search Icon Main Search Icon Video Play Icon Plus Icon Minus Icon Picture icon Hamburger Icon Close Icon Sorted

Community and Q&A

I have been reflecting on heat pumps in general lately

vensonata | Posted in Energy Efficiency and Durability on

After reading “Solar thermal is really, really dead,” which produced quite a reaction, is it possible that heat pumps themselves are at least wounded?

PV at $3.50 watt installed… Note that heat pumps are still great for all places where PV is impossible.

GBA Prime

Join the leading community of building science experts

Become a GBA Prime member and get instant access to the latest developments in green building, research, and reports from the field.


  1. Dana1 | | #1

    Heat pumps are only wounded when the cost of electricity (from PV or any other source) falls to such a ridiculously low level that the up-front cost of the heat pump relative to resistance heating is not viable on a lifecycle cost basis. At 3 US cents/kwh that may be compelling, but most of the world is paying 3-10x that much for electricity.

    The levelized lifecycle cost of grid-tied PV output is under 10 US cents for large scale arrays, but for rooftop solar in most of the world it is still over 10 cents/kwh, in some places much more.

    But the "learning curve" for grid tied PV is still pretty good, with a long term trend of dropping 20-25% in cost every time the installed base of PV doubles. With the prospect of seeing perovskite/silicon hybrid PV at an efficiency north of 30% being commercialized in the next 1-2 decades, the higher efficiency alone will reduce the "balance of system" costs dramatically (half the racking, half the labor). Perovskite thin film PV of various types should be VERY cheap, but panel costs with standard silicon technologies are already getting pretty cheap- the rest of the system is a bigger fraction of the installed cost than the panels themselves.

    There is at least a remote chance that a levelized cost of electricity of 3 cents/kwh will be realized with next-generation PV technology within the next 20 years, at which point heat pumps have to be really cheap or really efficient to compete in heating applications. Of course heat pumps will always have a place in cooling applications, but they don't necessarily have to use compressor technology. There are at least some newer reasonably efficient solid-state thermoelectric materials being commercialized, but they're not able to compete with refrigerant/compressor based technologies just yet.

  2. charlie_sullivan | | #2

    Dana captured most of the important considerations, but there is another factor here: The daily and seasonal variation that you impose on the grid. As more people use solar + electrically driven heat, whether it's heat pumps or resistors, there is going to be a big load on the grid on cold February nights, without any help from all of those solar panels. Sooner or later the structure of consumer electricity bills is going to require or permit being billed in a way that will make it attractive to reduce that winter night load from 10 kW with resistive heat to 4 kW with a heat pump.

    Once that shift happens, there will be other things that start looking more attractive again:
    A super insulated envelope not only makes your winter-night steady-state demand lower, but also increases the thermal time constant of your house, slowing down its response to outside temperature swings, allowing you to ride through a frigid night without a spike in electric consumption. A hydronic heat pump allows heat storage in a water tank to shift the timing of electric consumption according to the needs of the grid. Maybe we'll even see ground-source heat pumps start looking interesting again.

  3. vensonata | | #3

    Dana, good answer, let me just add a couple of numbers. Let us say at 15 cents kwh a mini split saves you 60%. So its like electricity at 6 cents kwh that you are heating with. Add in the cost of the heat pump over 12? years . Now you are at maybe 8 cents kwh. (These are just very rough guesses) So if Pv can produce for 8 cents kwh or less then thats it for heat pumps. At $3 watt over 30 years at 1500kwh per year = 45000kwh for $3000. That is 6.6cents kwh. Now in places with electricty at 10 cents kwh it is a slightly different story but still it seems close.
    I presume we agree that a cop of 3 is actually 66% reduction in electricity use and that 60% is a realistic expectation over the year in savings. The full price installed of HP is up for grabs since we have not specified models etc. I'd be happy to know if I'm off in these numbers.

  4. vensonata | | #4

    Reply to Charlie. Your answer is a big picture answer, the spirit of which I agree with. But many of those considerations about the grid capabilities are a long way down the road. If we think in terms of the next five years only, am I in the ballpark? By the way I have expanded my post above with a reply to Dana Dosrset.

  5. AlanB4 | | #5

    As long as the cost of electricity is above zero you will never lose money by using less electricity. I will co-opt the fixed cost going up argument right now, if part of the fixed cost is built into the usage rate then it will have to be shifted, but it can't cost more then the electricity one uses otherwise we would be paying most of the bill in fixed cost and the cost of electricity would be so low it would be an abstraction which it is not. In short if the cost of electricity fell by 90% and the fixed cost went up 50% you would still be 40% ahead (all numbers exaggerated to show effect).
    Your comparing apples to oranges, that article was saying the cost of solar electric hot water (produced by heat pump) has become lower then solar thermal produced hot water. This has no effect on heat pumps vs gas, oil, propane because these other commodities have not dropped in price below heat from the heat pump.

  6. vensonata | | #6

    There has been some great discussion here. It has forced me to further my inquiry into heat pumps vs PV. Briefly we should go back to the comparison between Heat pump water heaters with resistance water heaters. Lets forget about solar thermal since it was found to be more expensive. So with the addition of more PV the resistance heater and the heat pump were neck and neck at $3.74watt for pv. As pv price falls the advantage goes to the resistance heater. Does this apply to all air source heat pumps? Yes. The math is that although PV produces half the amount of energy for the same price as the heat pump, the PV lifespan is twice as long as the heat pump (at least twice as long). Now all this rests on the assumption of the grid price for the heat pump. The higher the grid price the more Pv has the advantage. At 15 cents kwh PV pulls away. The lower PV goes in price (and it is falling and can go to German and Australian prices nearly half of the U.S. rates) then heat pump cannot compare.
    Alan B's point about the percentage of saving always being the same with heat pump is right, but we have to add the price of the heat pump over its 12-15 year lifespan on to the cost of the electricity. For the first 5-6 years the heat pump doesn't save anything, only from year 6 to 12-15 does it reduce your bill by 60%. PV also doesn't save you anything until year ten (unless you got rebates etc to artificially lower the price) but for the next 20-25 years it saves at 100% i.e. it is "free" after it is paid for.
    So to conclude: exactly as the comparison of heat pump water heater to resistance water heater with additional solar panels found them close, so to does it apply to space heating.
    Another advantage to resistance electric heaters is their simplicity, low cost, and the ease of distribution throughout the house (such as bedrooms). I am not saying there is no use for heat pumps, they are brilliant solutions whenever space for PV is unavailable, which is probably 65% of all cases.

  7. GBA Editor
    Martin Holladay | | #7

    Your analysis makes sense, but it depends on net-metering agreements that credit homeowners with the retail price of electricity.

    There is a fly in the ointment: some U.S. states are undermining net-metering agreements, piling new fees on PV owners. Until batteries get a little cheaper, homeowners in these PV-hostile states are between a rock and a hard place. For more info, see Wisconsin Alters Net-Metering Rules.

    And even good batteries won't let PV owners use electricity generated in August to stay warm in January.

  8. vensonata | | #8

    Yes Martin, I must admit the variables concerning net metering are beyond anybody's ability to calculate. It seems that those who are building net zero houses tend to use both PV and heatpumps, so that is another layer of math to ponder.
    This discussion reminds me also of the debate over PV tracking devices which increase the production over stationary panels by up to 50% in summer. Alas it turned out as PV fell in price it was cheaper to just buy more PV... the trackers couldn't pay for themselves except when PV was at $12 watt! So these are the historical paradigms that I have been applying to some of our newer energy saving devices such as heat pumps.

  9. vensonata | | #9

    sorry...duplicate reply.

  10. Expert Member
    Dana Dorsett | | #10

    One has to look at the cost of money when evaluating these investments, and the risk of potential repairs, maintenance. Net present value calculations of the investments are really the correct starting point.

    Viewing it as "saving" some percentage is not the right way to view it, since that presumes some paradigm for the cost of energy to which it is being compared. The total cost of the energy, including the total lifecycle cost of the heat pumps and PV need to be factored in. When PV get's cheap enough, the up-front cost of the heat pump may not be a cheaper than simply more PV.

    So far only one US utility is considering applying "demand charges" to residential bills rather than using block rate fees based on energy use for paying for the grid infrastructure, and that is a small municipal utility in Arizona. Demand charges are based on the highest-use half-hour or hour of energy draw from the grid, which is a better measure of how much grid infrastructure is required to support a custmer than a fixed amount per kwh of energy use. If this becomes the new paradigm for residential rate structures in the age of ubiquitous distributed energy, that will improve the economics of heat pumps at the same time that it DESTROYS the economics of electric tankless hot water heaters and other heavy draw intermittent loads.

  11. vensonata | | #11

    Dana, yes there will be many a strategy on the part of utilities. Remember, they have never had to compete, and their entire model of business is like someone who has inherited a lot of money, it wasn't beause they were clever or worked hard. For every move they can make it is relatively simple and reasonably economical (not cheap!) to make a counter move on the part of the individual. Low cost PV is here to stay and heading down. Home batteries as in Germany now will allow some peak shaving and some storage, but the grid will be forced to cooperate or the game, as they say, "is on!" And such clever gentleman as are found on GBA will make life difficult for anonymous giants who are entirely too comfortable.

  12. Expert Member
    Dana Dorsett | | #12

    I understand quite a bit about the variations on and the history of utility business models, and they vary by quite a bit in the US. Some utilities (eg Georgia Power) are large state-wide monopolies that own both the grid and the vast majority of the generating resources, and they get compensated by a guaranteed return on their capital investment by rate-basing all of their expenses. Others are "decoupled", even barred from owning over a certain percentage of the generating capacity on their grid, and compensated on reliability and judicious grid infrastructure capital investments related to reliability, and the energy costs (often power purchase contracts from multiple independent generators) are passed through without mark-up in to the rate. But what's going on right now in NY is radically different, turning the grid into basically an equal access network for willing buyers & sellers of power.

    Charging residential customers a demand-charge to cover the costs of the grid infrastructure independently of energy use may be viewed as some as a protectionist practice, but in fact it's a fairer assessment of grid infrastructure costs than has been traditionally been done, proportional to energy use. It's really only protectionist/anti-distributed generation if they only apply demand charges to customers who are also power generators. With the standard block fixed rate type structures power-sippers who use the same amount of energy as power gulpers pay the same amount, but the guy with the heat pump water heater and 2-3 tons of modulating mini-split only needs a small fraction of the grid capacity/infrastructure of the customer with the electric tankless hot water heater and 8 ton air conditioner who uses the same amount of total energy. It's a serious cross-subsidy of power-sippers toward power-gulpers, since it's the gulpers that define just how much capacity the distribution lines transformers and substations need to be, whereas the power sipper is paying for grid capacity that they don't actually need or use.

    This is exactly the problem that has driven electricity prices stratospheric in Australia, where the grid capacity was built out to insanely high multiples of what's actually needed, on the presumption that power demand would continue to grow forever. But higher efficiency equipment flattened the demand curve for a decade or more (even as they were continuing to build it out), and as distributed PV became ubiquitous, grid power use has been slowly shrinking, not growing. As the prices rise to pay for the over-built grid with lower kwh sales, the financial case for rooftop PV on your side of the meter goes up.

    Battery storage on the customer's side of the meter is allowed in Australia, and SolarCity is poised to take a big piece of that market, since excess power put onto the grid is compensated only at the wholesale energy rate, not the full retail rate the way is it done in most of the US. Last year in California battery storage on the customer's side of the meter became legal (and regulated) just last year, driven in large part by SolarCity (the largest residential PV installer in the US.) But that is not the case everywhere in the US (yet.)

    Some utilities have heard the bell and are very supportive of distributed power, and are adjusting their business models to accomodate. (The local utility in Austin TX, is one, the largest utility in the state of Vermont is another). Hawaii's investor owned utilities didn't wake up and smell the coffee burning until it was boiling over, and was recently acquired by a large company with distributed power experience to avoid abject bankruptcy failure (this, only 3-4 years after rejecting a buy-out that would have been several times larger by a consortium of Wall Street investors.) Many utilities are tacking on fees to PV owners to "pay their fair share for the grid", but that is at best a delaying tactic. PV (and storage) is getting cheap enough fast enough that demand for power purchased from the local grid utility will continue to shrink for the next few decades, and by 2040 outright grid defection will be rational for those with sufficient access to both sun and capital, if they keep jacking on fees for solar customers.

    From a regulatory point of view a handful of US states are in the forefront on this, NY, MA, HI, CA to name a few, whereas others are still playing push-back mode. Of the latter, AZ is one of the biggest ongoing mud-fights, with vertically integrated utilities struggling to avoid a serious stranded asset problem. As has been happening in Germany, utilities that fight back too long without writing down some of their no-longer needed assets and adjusting their business models WILL fail financially.

  13. vensonata | | #13

    Dana, thank you for the nuanced discussion of utilities and the grid, it is a big topic. One follow up comment regarding heat pumps vs Pv. The source of the electricity for the heat pump may be very dirty, though it reduces ones use in quantity. Pv does not require the source and may in the sunny season actually reduce the source electricity by back feeding into the grid. This factor alone might sway some to pv over heat pump, though there may also be an argument to use both in combination.
    One final note. In such places as Arizona where the winter solar fraction is 70% of summer it may now be possible to heat entirely with Pv and of course cooling in their dreadfully hot summers can be done with the very same array. Some thermal storage may be part of the formula. Again it is only possible in certain areas at present and only if the array space is available.

  14. jpkidd | | #14

    Just to agree with most every comment, PV and resistance heat may work for the individual. But in midwinter where I live, solar is 30% of summer intensity for 8 - 10 hours of a blue sky day - some one is going to be storing or spinning no matter how cheap PV gets! Let's say we've (well, not me) solved the supply side with PV, resistance heat loads up the storage side of the PV solution, as does any less efficient than possible technology or usage of electricity. In my climate people reject thermal mass building solutions but I guess as PV's cost lowers, the range for thermal mass solutions creeps northward, just like cold climate air source.

  15. Expert Member
    Dana Dorsett | | #15

    Another nuanced discussion on the utility vs. distributed PV generator vs. non-solar ratepayer situation showed up on today's greentechmedia blog (from the perspective of one of the SolarCity family-insiders):

  16. vensonata | | #16

    Reply to Peter Kidd, I should make it clear that in most places in the northern states pv is out of the question for providing direct space heat in winter. The question really is. "should I spend $5000 on a heat pump which will last 15 years or $10,000 on pv which will last 30 years?". Which will reduce my electrical demand most? Now this presumes you have a feed in tariff with the grid. At costs which are soon to appear, of $3 watt installed Pv is the better investment. You won't heat with it but net zero houses over produce in summer and get it back in winter. Some net zero houses use both heat pump and pv there any point to that rather than pure pv and simple cheap electric resistance heaters? That is what I am getting at. Now of course you need more roof space, but if available, pv seems to be the simpler solution.

  17. Expert Member
    Dana Dorsett | | #17

    A feed-in-tariff (FIT) approach has never been taken anywhere in the US that I'm aware of. Most utilities to date have opted for "net metering" at the residential retail rate, effectively running the meter backwards whenever the PV is delivering more power than the house is using. Any excess power to the grid at the end of the year is usually a "gift" to the utility, reset to zero on an annual basis. (Some utilities are trying to get approval to zero the net metered excess on a monthly basis.) This is very different from how PV was compensated in most of Europe (and Australia), where power exported to the grid was paid a premium in the form of a FIT, usually well above what would have been charged to the customer for having used that much power. High FITs for PV installed 8-10 years ago in Germany are 2x the average residential retail rate for power in Germany.

    Once fairly substantial, FITs in Australia are now pretty much gone for new installations. Worse yet, many utilities do not net meter at all, paying zero for power exported to the grid. Others net meter, but pay only the wholesale cost of energy for the instantaneous excess power exported to the grid. You could be net zero energy for the day/week/month, and still have a hefty bill, since power drawn when your PV was puttting out less than your entire load costs 3-4x what you are paid for what went on the grid. But there the per-kwh bill for just the grid-use charges are about 1.5x the average retail (grid + power) charges in the US. This imbalanced net metering plus the high cost of the grid makes battery storage on the PV customer's side of the meter very attractive, since the power they store & use is power that would have been compensated at a very low rate had it gone onto the grid. Some PV owners in Australia have installed systems that sense & prevent the instantaneous net export to the grid, shunting the "extra" PV output to a heating element in the hot water heater tank, since heating hot water with grid power would be 3x or more what they would have been paid by the utility.

    The economics of heat pump vs. PV have to be analyzed based on the local FIT or net-metering environment but with simple net-metered PV (not FIT compensated) the heat energy leveraged by the heat pump is still almost always cheaper on a lifecycle basis than the cost of PV power on a lifecycle basis. When the installed cost of PV drops under a buck-a-watt (hope springs eternal ! :-) ) this may no longer be the case. But right now it's not even close in the US, where the average installed cost of small scale PV is still ~2x what it costs in Germany (where PV output is only marginally worse than much of the US due to weather & latitude) or Australia (where the PV output is subtantially higher than the US average.)

  18. GBA Editor
    Martin Holladay | | #18

    For a while, a few utilities in Florida were offering feed-in tariffs with above-retail reimbursement for PV. Vermont also has a 19 cents/kWh feed-in tariff for some customers with PV. These contracts are limited, however, and not available to all customers.

  19. mackstann | | #19

    Oregon had a feed-in tariff of 39 cents. You had to be selected via lottery. It may still be running but I think they shut it down.

  20. vensonata | | #20

    Dana, yes the uncertainty of our relationship with the utilities is certainly a sticking point. By the way Ontario, with its population of 14 million, has a residential FIT of 70 cents kwh! Guaranteed for 20 yearsIi believe. Amazing. But yes we really need to do the math on simply cranking the meter backwards....and especially when they zero the excess in their favor at the end of the year. So possibly pv has more advantage in hot climates than cold. It corresponds to maximim useable production. Although as I mentioned before that Arizona is almost perfectly suited for optimal use and solar production in both winter and summer. Certainly Alaska is a killer for both heat pump and pv, since it is so dark and so cold. The sunny mid west is an interesting toss up..sunny but very cold with reduced efficiency of heat pump. We need to run this through a rocket scientists' brain to get the final formula, I think!

  21. Expert Member
    Dana Dorsett | | #21

    I'm surprised at the number of US utilities that drove the FIT route, (even if only in a limited fashion!)

    I suppose Austin's Value of Solar Tariff (VOST) is also truly a feed in tariff too, so I guess I was talking out of school on that one (I knew about it, but wasn't thinking about it when typing away.) Thanks for the education Nick, Martin!

    A bit of web searching found that CA had a FIT that only applied to the three largest investor owned utilities that began in 2008. It was adjusted by market- whatever was necessary for those utilities to get their mandated share installed, up to a cap. The initial cap was met then raised in 2012 to a total of 1GW total installed base among the three utilities, after which it expired.

    I'm also surprised that the FIT in Ontario is still that rich. IIRC Australia's FIT peaked at AU$0.44, but Germany's started out significantly higher, but has been falling in calendar-linked stages. If there's a "right" way to do a FIT, (or other subsidies), it would be to base it on steps in the size of the installed base rather than the calendar (sort of CA-style.) Germany got caught hard in a PV deflation trap, where the price of PV fell MUCH more rapidly than the FIT for several years. That created a PV rush which got a lot of PV installed quickly, but the ratepayers at large paid way too much for the PV relative to it's actual cost to the PV owners. Then when the FIT got suddenly slashed in response there was a PV bust, putting the local solar industry back on their heels.

    The Ontario FIT is about 4x what I've been paying for electricity, which means there is no way an air source heat pump would pay as well as extra PV, if they pay the excess over net. But the math is still a bit odd. If you can free up more of your PV output with a heat pump to sell more power at 77 cents/kwh, even at a lousy COP of 1.5 it may still be a significantly cash-positive investment.

    Both PV & heat pumps have advantages wherever grid power is expensive, and the regulators don't allow the utility to pick your pocket on PV. The annual insolation in AZ isn't even 2x what it is in southern Ontario, so it's not as if they would be at a great advantage in a net-metered environment. Their residential power rates in AZ are also a bit higher than most Canadian utilities charge, which is the real cost advantage. But the utilities in AZ are even fighting straight net-metering, often playing sneaky & dirty, let alone paying a FIT. This is just their latest backhanded attack on their solar competitors:

    There is no single rocket-science formula, since the regulations, power prices, and FITs are constantly changing. The lifecycle cost of rooftop solar at New England installation costs & insolation levels is on the order of 12-14 cents/kwh. which is about half the wintertime residential rates for much of New England this year, but only slightly below the recent 5 year averages. Even at 13-14 cents/kwh a ductless mini-split heat pump is still a great investment.compared to resistance heating. But at 5 cents or lower it gets pretty squishy, harder to make the case for the substantial up-front cost. When PV hit's a buck a watt installed price, the lifecycle cost will be in that squishy territory that requires a sharp pencil to know for sure.

  22. wjrobinson | | #22

    Superinsulate #1

    Orientate your home properly #2

    Use mini splits #3


    End of story.

    PV is getting very very dicey and will get worse as to when Scott Walker is our new president. Just mentioning PV will get you burned at the stake let alone having panels in one's possession.

  23. vensonata | | #23

    I am back with a new way to look at comparing heat pumps and pv. I hope this thread is not dead yet. Anyway, we have all been off the track by bringing in the grid at all into this discussion. And in fact if you follow my argument it looks like PV is about twice as cheap as a heat pump! By the way I am not hoping for either side in this comparison, may the best system win.
    So let us start with a house near, say, Boston. It is a pretty good house and requires 10 million Btu heat per year. Let us install a Heat pump and some resistance heaters for back up for a modest $5500. Now we have cut down our heat demand by 60% or 6 million Btu. Lets convert that to Kwh...that is 1764 kwh that we have saved. How much PV does it take in Boston to produce 1764kwh per year? I checked on PVwatts and one kw array produces 1350 kwh per year. So lets say we put up 1.5 kw array, that gives us 2025 kwh per year. That is more than our heat pump saves. So how much does the PV cost? At $3.74 per watt cost about $5600. Wow, the two system are close indeed! Except...the PV lasts twice as long as the heat pump. Voila, PV is easily half the cost of heat pump over the life of both systems.
    Notice the grid price is irrelevant. It is the same for both systems. As well Pv only needs to make the meter spin backwards it does not need any payment tariff.
    This conclusion startled me, I really didn't expect it. But if it is true then it sure makes a difference economically with a great advantage going to PV with cheap resistance electric heaters.

  24. wjrobinson | | #24

    Ven, Problem. What to do about the Scott Walker types that want us to stop selling PV back summers for winter repurchase? Houston, we got a problem as this looks like a trend coming.

  25. user-626934 | | #25

    I wouldn't bet too heavily on net-metering as we know it lasting much longer than five more years in high PV penetration locations. Grid dynamics and the big picture are an essential part of a proper discussion here on GREEN Building Advisor.

  26. GBA Editor
    Martin Holladay | | #26

    I agree that "Grid dynamics and the big picture are an essential part of a proper discussion here on Green Building Advisor."

    For more on this issue, see my latest blog: The Evolution of Superinsulation.

  27. vensonata | | #27

    Yes, grid dynamics are very important, also, rather uncertain. That is why the simple comparison of PV and heat pump should be based only on what we know exists today. And as it stands Pv with grid connection has a great advantage over heat pump.
    The next part of the discussion is: "can pv be used directly without battery for your house needs so there is no entanglement with your grid connection?" The answer is... probably yes, with some conditions limiting the amount of Pv production so one does not waste electricity with over-production in Summer. But this deserves a fresh discussion forum. Perhaps someone will offer up the question which opens this new can of worms!

  28. Dana1 | | #28

    Ven: It takes a much higher efficiency house than a pretty-good-house (PGH) to end up with a heat load of only 10MMBTU/year. That's 100 therms- less than half of what most would spend on hot water. In a PGH the annual energy use for space conditioning & hot water are about the same, IF leveraged with heat pumps. The cost of a building envelope with U-factors that low is likely to be well beyond financially rational when compared against the cost of PV @ $3.74/watt. That is the basis of pretty-good-house or net-zero house vs. PassiveHouse discussion- even on 100 year lifecycle basis for the insulation, the additional cost of the "extra" insulation doesn't pencil out favorably against covering the additional load with PV.(leveraged with heat pumps or not.)

    In a Boston climate, a better-class mini-split will deliver more than 3x as much heat per kwh as resistance heating, and the net present value of PGH against something closer to code min would have to be updated as well. For "extra" PV + resistance heating to compete against a mini-split in a straight net-metered environment the PV has to produce more than 3x the kwh per $ of investment over the lifecycles of the PV or heat pump. That's independent of the annual load, whether 10 MMBTU/yr (PassiveHouse) or 1000 MMBTU/yr (a leaky sub-code wreck).

    The grid-stabilization and infrastructure load reduction value of distributed solar on the customer's side of the meter has a higher value to other non-solar ratepayers than the residential retail price of electricity, at least until the PV peak output on the load side of the substations is above 100% of the instantaneous load, as is happening on some local grids on Oahu. It doesn't take a lot of distributed battery to solve that problem though, and in PV saturated neighborhoods I expect local storage may become a requirement for any new PV, unless grid-sensitive electric car chargers show up first. Under New York's new regulations it looks like utilities would be allowed to incentives PV &/or car-chargers at different rates in different locations on their grid whenever the subsidy would cost less than utility-owned grid upgrades to deal with the substation capacity or local grid stability problems. It would be a travesty to have widespread grid defection, since it requires MORE infrastructure investment (on the grid-defector's part) than it would for the grid operator to manage it, AND it would deny the rest of the utility customers the capacity & stabilization benefits of the distributed resources.

    A primary impediment to financial rationality in this situation in the rate structures that evolved out of the old-school grid models, that promoted infrastructure investment by regulated monopoly utilities for both electricity production and distribution. There is still a case for incentivizing maintaining the grid, but not so much for expanding it, and the cost of actually generating electricity with PV is falling fast, and will be at a lower lifecycle cost than the wholesale cost of centralized fossil & nuke generators before 2030, which makes widely distributed PV a financially rational way to power the grid. A lot of very smart people (in NY and elsewhere) are spending a lot of time & energy re-thinking how to rationalize rate structures to make this happen, which would make electricity cheaper for everyone on the grid (not just the PV owner/operator). Case in point:

    There will be bumps on the road as the rules & business models change, and as utilities that have overspent on peak generation or grid infrastructure attempt to avoid stranded assets and write-downs, but the notion that it will take widespread grid defection to force the necessary changes doesn't seem likely.

  29. charlie_sullivan | | #29

    Ven says, "can pv be used directly without battery for your house needs so there is no entanglement with your grid connection?"

    Then Dana argues convincingly that that's not the right question but I'll answer the question anyway. Right now a modulating mini-split heat pump modulates based on the heat demand. It's would be possible to modulate based instead on the power output of a PV system. That would require new control boards in the minisplit and in the PV inverter(s), developed in cooperation between those manufacturers. Not a DIY project--for now it's just a thought experiment.

    With a 20% efficient PV panel (high end) and a COP = 3 heat pump, the net efficiency is 60%. That's pretty similar to a hydronic solar thermal collector.

    So does that make any sense? Well, using solar thermal hydronic collectors for heating lost favor a while ago. the main advantage of it vs. passive solar (using solar gain through windows for heating) is that you can use a big tank to save up heat from a sunny day for a cold cloudy day. If you disallow batteries with the PV + heat pump system, you are back to only heating when its sunny.

    I expect that in the future we will want the PV on a given building to talk to that building's heat pump and use do more heating when the sun is out, but I don't think that will be the whole story. Both will also talk to the grid and through the wider diversity of sources and uses on the grid we'll be able to make it through cold cloudy days by curtailing other electricity use and by ramping up other electric power sources on a grid-wide basis.

  30. Dana1 | | #30

    The cost, maintenance and lifecycle of active solar thermal is different from PV and mini-splits. Even at equal solar efficiency it takes more square meters of panel to heat a place with solar thermal than in a net-metered PV + heat pump case. Off grid PV + heat pump adds a huge cost factor in battery size, and is completely uneconomic, but then so is off-grid active solar.

  31. vensonata | | #31

    Dana, 10 million Btu would be passivhaus at about 1800 sq ft. PGH at say, 1200 sq ft. So although it is just a convenient number to work with, its not unrealistic. Domestic hot water, from the old discussion on solar thermal vs pv came in realistically at about 8 million Btu year. That these two figures, heat and hot water, are similar, is to be expected in a very good house. The other phrase we have to be careful about is "3 times as much" referring to heat pumps at cop3. We should be saying "60% less than electric resistance" That keeps us from making mistakes with the math. Cop 2 is "twice as much", cop 3, is only 16% more efficient than cop 2.
    Anyway, the real issue lies with the use of PV divorced from the grid connection, which we actually want to keep. We don't want to heat with PV in winter of course, though we can use up our 2000kwh production on our domestic hot water directly, 8 months per year. Ultimately it produces a better economic investment than ASHP heating our house in the worst time of year. We could also find other direct uses for pv electricity without storing it in batteries, but thermal storage is the easiest. There is no need to feed back into the grid...separate circuits for PV. It is also feasible to use a very small battery of say 4 kwh ($500) to smooth our day use. I am presuming that the non heating electrical day use is perhaps 5 kwh at most. That, by the way, is plenty if you are the least bit conscious.
    A good part of this discussion is due to the fact that grids, such as Arizona, are obviously at risk because of the perfect solar conditions for PV, and being reactionary types they make impulsive protective rulings which just infuriate everybody who thinks PV is the future.

  32. Dana1 | | #32

    PassiveHouse levels of energy use aren't "realistic" when looking solely at net-present-value of future energy use savings, unless you project long term energy price inflation. (The falling cost of PV indicates that we may in fact be looking at long term energy price DEflation, over the lifecycle of a house.)

    Resistance heating: 3412 BTU per kwh

    Pretty-good mini-split in Boston: 10,,236 BTU/kwh

    Calculate the lifecycle cost of the 6824 BTU difference when produced by PV, against the lifecycle cost of the same 6824 BTU when provided by a mini-split. At the current cost of PV in Boston, the mini-split is still winning (for now.)

    Heat pump water heaters still win too.

    Mind you, if the mini-split is only being asked to deliver 10 MMBTU/yr it's lifecycle is probably going to be substantially longer than when operated at the duty-cycle need to heat a "Pretty Good House" (using the definitions bandied about by Martin Holladay ,et al.) or IRC 2012 code min house with 3-8x the annual load.

  33. vensonata | | #33

    A final update on whether my numbers for heating in Massachusetts are realistic. They are! Please check out Marc Rosenbaums complete detailed stats on his 8 house net zero community in Mass. The heating figures are almost identical to what I plucked out of my head. His heat pumps performed at cop 2.25. And resistance heaters were also required. His heat pumps cut the bill by about 1700 kwh per year which is pretty well identical to my estimates. Since he has 5 kw pv arrays on the roofs carefully monitored and grid tied you can see their actual production is 7000 kwh per year or 1400 kwh per kw of array, slighty better than my guess of 1350kwh. So, somewhat to my own surprise all my figures are very close to actual monitored houses. The only thing left to say is that except for the precarious future of grid connected pv, I feel I can rest my case for pv being half the cost of heat pumps over their lifetime. By the way Marc Rosenbaums stats appeared right here on GBA but the whole beautiful pdf is on google as well.

  34. iLikeDirt | | #34

    Things are a little different I think if you leave a heating-only climate. Once you go into a climate with cold winters and warm summers, you can consider that the heat pump represents a free air conditioner. Likewise in places that are hot most of the time with mild heating loads (like the American south) where the heat pump can become a heater when needed so you don't have to install a whole separate system. And in places where you're both heating and cooling a lot (the midwest, the rocky mountain states, the middle atlantic, the high desert, etc), the same is even more true.

  35. vensonata | | #35

    Nathaniel G. I guess the formula for valunig a heat pump is: the more it runs the more it saves you. In a warm climate it doesn't run as much in the winter but it runs more in the summer. Add together and compute to total kwh saved for the year. Take that number and compute to replace those kwh with PV. If the price is equal PV is half the cost, because twice the lifespan. Now, down south the pv produces when cooling is needed, that offsets the cooling demand in real time. Take Arizona or New Mexico or Southern California. The pV payback is much higher (up to 50%) because of the number of hours per year of sunshine. So the benefits of PV might be even greater, in fact, you can calculate 1800kwh year for 1 kw of pv in those regiions. for $3500 installed. Try to get 1800 kwh savings from a Heat pump that costs $3500 installed. Maybe you still has only half the lifespan of PV and is therefor twice as expensive.

  36. iLikeDirt | | #36

    My point is that a heat pump is not "saving" you any kWh for cooling, since there's no electric resistance equivalent of an air conditioner. If you need cooling, you need a refrigerant pump machine of some type and a mini-split heat pump is simply one of many choices (unless you live here in the New Mexico high desert where a swamp cooler works fine, but I digress). So the heat pump will only save you kWh for cooling if you're comparing it to an old low-SEER clunker AC or something. But I guess I get your point: if you were contemplating replacing a SEER 10 unit with a SEER 20 mini-split heat pump, the projected kWh savings would have to be compared against the kWh produced by adding more PV.

  37. vensonata | | #37

    Nathaniel, Yes, I neglected to provide an alternative to the heat pump as cooler. If the cooling load was significantly larger than the heating load you would have to oversize the heat pump for winter when it actually has cop of 3. In summer I don't think it is any more efficient than a comparable quality air conditioner. And perhaps a Coolerado would outperform them both. And then there is ice storage...apparently the best of all and perfect for direct use from PV. (By the way I'd love to hear of anyone using PV directly to charge their "ice battery"! )
    Well now that we have opened the subject. Since Pv grid tie is now on shaky ground according to some, is it not a good use of PV to use water storage for domestic hot water and some shoulder season space heat and ice battery in Summer? Certainly perfect in the sunnier climes of NM, AZ, CA etc.

  38. iLikeDirt | | #38

    My swamp cooler certainly outperforms the highest SEER AC or mini-split heat pump on the market. For 200 watts and 6 gallons of water an hour (much of it re-used for irrigation), it keeps my house at 72-76 degrees for almost the entire cooling season. And it is so simple that there's no reason why it shouldn't last for 30 years or more; every part is homeowner repairable or replaceable. The more advanced units with two-stage designs and/or that don't add humidity should be no-brainers for anywhere but places with super-humid summers, IMHO.

  39. vensonata | | #39

    I am back with some new numbers....and it is bad news for ASHP, unfortunately. I really like these things but it turns out I am irrational if it is for economical reasons. After coming across Marc Rosenbaum's fabulous data on his 8 house net zero community in New England, all kinds of revelations unfold out of that document. What is nice is that it involves both PV and air source heat pumps, and the heat pumps are high quality Daikin, and they are measured in actual use not just manufacturers claims.
    Now one caveat: the numbers are very specific to a modest sized high performance house. They will change if applied to a 2000 sq ft rancher from the 80's. SO here goes. The kwh savings per year were approximately 1700. To be clear, the electric bill was 1700 kwh lower than it would have been with only electric resistance heat. I do not know the exact cost installed of the heat pumps but I am guessing around $5000. If the kwh grid price in Mass. is 15 cents then the savings per year is $255. Oh, oh! It takes 20 years to pay off the heat pump! But the heat pump died sometime around year 15-17. Conclusion, heat pump is a net loss over resistance electric heater! What about PV? It had similar cost to the heat pump for the same 1700 kwh/year production. It is paid off at year 20 and then produces free electricity for the next 10-20 years.
    Well I don't know what to think except I definitely would not put a heat pump into those houses...and probably would require electric rates of 25 cents and a cheaper heat pump to make any sense at all.

  40. STEPHEN SHEEHY | | #40

    Ven: if it helps your analysis any, I'm paying $3200 for installed Fujitsu heat pumps in Maine this year and $3.30 per watt for PV. Both prices do not include credits or rebates (30% federal tax credit for PV, $300 rebate for heat pump from Efficiency Maine.)

    Our electric rate was about 15¢ per kwh, but is going up next month to about 19¢.

  41. vensonata | | #41

    Stephen Sheehy, Thanks, incredibly helpful. Those are the three numbers that tell much of the story. If your heat pump saves you 1700 kwh/year, then a 1.2 kw array will produce that much per year. It will cost $4000. Your heat pump costs $3200 and will last (lets hope) 17 years. At 15 cents kwh it will save you $4335 over it's lifetime, a total net gain of $1135! congratulations. Over the same period your PV will save you $4335 and total net gain of $335! again, congratulations. However at this point the tortoise pulls ahead of the rabbit and for the next 17 years your PV will earn you $4335! (if, electricity is still at 15 cents kwh!) So we see that both systems are good investments but pv, in the long run is almost double the return.
    Warning: I am not a mathematician, I am a Buddhist monk, so you would all be advised to get a second opinion on these numbers!
    Lets run the numbers with your rise in electric cost to 19 cents kwh and with your rebates.
    Your heat pump pays for itself at 8.9 years. and over the next 8 years of its life earns you $2,591. That is based on $2900 price of heat pump after rebate. Really looking like a good investment with those figures!
    The Pv pays for itself after 8.6 years and by year 17 earns you $2,691. Then over the remaining 17 years of its life it earns a further $5491. A very, very good investment.
    Conclusion: So it seems heat pumps are a good investment at the right figures. And Pv at these figures is a spectacular investment!

  42. STEPHEN SHEEHY | | #42

    Ven-A 1.2 KW array doesn't keep me warm. Don't forget that if one eliminates the heat pump, one still needs something to turn the PV power into heat. If one uses electric resistance heat, that costs something to install. So to compare the two options, heat pump v. PV, we need a number for the installed cost of electric heat. Resistance heat is cheap, but not free to install.

    One final point: here in Maine, a 1.2 kw PV array will produce an average of about 1450kwh. This winter, probably not so much, given that it snows every two days.

    This discussion is really useful. Not every decision concerning construction should be based on dollars, but every significant one should at least have the financial impact available to aid in the decision.

  43. vensonata | | #43

    Stephen Sheehy, In the case of Marc Rosenbaums houses back up resistance heat was also installed! And in all 8 houses it was used intermitently. There is an article by Marc in todays GBA about what happens to heat pumps during snow storms...they struggle. So probably you are going to need resistance electric anyway. As to the Pv ratings, Pv watts is really quite reliable in its predictions and note that the derating for system loss as been reduced with new panels and inverter systems. Check carefully for optimal angle for year round net gain (usually latititude is best).
    Indeed, I am for exploring every nook and cranny of this comparison since many, many people will have to make this choice in the near future and the numbers are changing all the time. Advice that was valid even one year ago may already have passed its best before date.

  44. Airithol | | #44

    Thanks for the interesting discussion, all.

    Ven, your argument is becoming increasingly persuasive. As PV prices continue to drop the case for heat pumps becomes less clear. Didn't we just start getting excited about these things in the past 5 years (in North America anyways.)?

    That said, if PV can innovate and become more affordable, so too can heat pumps. They also double as air conditioners. My takeaway is they are no longer a 'no brainer' but deserve closer analysis. Perhaps they always did.

    I have read many forward looking articles on PV, energy storage and electrical grids in the past few years. It's a rapidly changing landscape, but what I've come to realize is that what is really happening is our perception of energy is changing. The tech just follows the sentiment.

    As I sit here and watch my 1 year old daughter play on the floor, I realize she will read about these past few years and the next 5 (10?) in school. Probably in history class under the title 'The solar (energy?) revolution'. Exciting times.


  45. vensonata | | #45

    Jason, yes as I inquire more I realize, at least as far as I have found, that careful consideration about Heat pumps is scarce. Unfortunately my math and accounting skills are limited, so I try to tread carefully out of my area of expertise. But while I have been tip toeing around I have blundered into what I believe is another major miscalculation on the most authoritative solar calculation site in the U.S. "PVWATTS". It is a goldmine of instant calculations about PV costs and production throughout North America. But as I did the most recent calculations on Stephen Sheehy's house (see above) I suddenly realized that at 19 cents kwh grid price both PV and heat pump are paid off in less that 9 years! That means that the calculation on PVWATTS are wrong because they add an automatic economic factor for the "levelized cost of power" (fancy talk indeed) of 7.5% loan over 25 years. So two big mistakes here: 1) the loan is paid off in 9 years. 2) PV panels are warantied for 25 years but they can easily last 40. And I suppose a third point is interest rates are lower than 7.5% these days, especially for a 9 year period. So this throws their numbers for a loop!
    What do I know, I am not a banker! But I would compute numbers based on the "displacement" of grid electricity vs trying to figure cost installed divided by lifetime production plus loan interest etc. So I am sure many of the people on GBA have much greater fluency with these formulas than I, and I would love to hear some competent math on this topic of "what is the real cost of PV?"

  46. Expert Member
    Dana Dorsett | | #46

    Try NREL's quick'n'dirty online LCOE calculator, at whatever price point or discount rate you like:

    For a New England type capacity factor for PV figure 13-15%. Retail the PV costs about $3.50, but the 3rd party ownership solar companies cost basis is trending rapidly to under $2 (before any subsidies are applied.)

    Panels can go 25+ years, inverters generally don't- figure 15 years between inverter replacements, 20 years best-case. Kick in at $20- $30/ per kw of array per year into the fixed-cost operations & maintenance part of the calculation to cover inverter swaps and other lighter duty repair.

    In a 25 year analysis assuming a 15% capacity factor, at a 3% discount rate and an un-subsidized $3.50/watt ($3500/kw) with $20/year per kw to cover the inverter swaps it comes in at about 17 cents/kwh. At 2.5% discount rate it comes in at about 16 cents.

    In the US there is (for the time being) a 30% tax credit subsidy, which would reduce that to $2.50/watt, at which point it comes in at about 12-13 cents/kwh using those low discount rate numbers. Bump the discount rate to 5% and you're back up at 15 cents/kwh.

    SolarCity will be (or maybe already is?) offering 4%/30 year financing (and 30 year monitoring, maintenance & warranty) for residential customers with good credit, and as the #1 installer in the US they are competitive on $/watt basis. It's pretty easy to beat New England standard grid-rates with that... long as the utilities don't bite back with excessive charges for grid use. So far that isn't the norm in New England, and the regulators may not let them get away with it, the way it has happened in some other regions of the US. (The utilities generally lose those fights, but there are exceptions.)

    The EIAs LCOE comparison numbers live here:

    note: The $118.6/Mwh (= $0.186 per kwh) LCOE for utility scale PV is grossly overstated. There are examples in 2014 of long term power purchase agreements for less than 1/3 that amount, both in Texas and Georgia. Most of the 3rd party leasing contracts for residential rooftop solar in my neighborhood are in the 16 cents/kwh range, which includes significant profit margin for the solar company beyond their financing costs. The cost basis of utility scale solar is much lower than residential rooftop. The world price is now under $2/watt (installed) for large arrays.


  47. vensonata | | #47

    Dana, Thanks for that. I notice that a new commercial array in Utah, somehow manages to sell to the utiility at 6.5cents kwh. Now I presume they make a profit on that, they are not a charity. And then surely they had to buy land for the array, and while commercial is cheaper to install it just can't be that much cheaper than a free rooftop and no transmission lines. As well, micro inverters are now warrantied pre installed for the same 25 years as the panels. With a tax rebate of 30% we are talking maybe $2.20 watt installed. As you pointed out the math for the NREL just cannot be realistic if commercial is sellling for 6.5cents kwh. I end up with math that suggests that the real price for residential at $2.20 installed must be closer to 8cents kwh. Somebody with a degree in freakonomics needs to crunch these numbers. I smell distortion somewhere in the government figures...and it wouldn't be so surprising if that was the case, would it?

  48. wjrobinson | | #48

    Ven, big commercial PV may be cutting the costs via the CO2 trading market?

  49. Dana1 | | #49

    The large arrays still get tax credits & other subsidies, and are cheaper per watt to build due to the economies of scale. SolarCity's overall cost basis is $2.09/watt as recently reported on greentechmedia:

    That's a mix of project sizes from 1kw to several megawatts, with most of it being in the sub 10kw range. That's their cost, before any subsidies. For the large arrays they're probably under a buck-fifty, but still over a buck.

    Note, the capacity factors of PV in Utah or Texas are also significantly higher than in New England. More annual power per rated panel-watt translates into cheaper power sales.

    EIA numbers regarding renewables has a history of using lots of outdated data, and late reporting. It's more due to lack of focus than incompetence or an intent to deceive. When the numbers are slewing as fast as they are with solar and wind they nearly always get it wrong, both on the cost, but also on the forecasted rate of implementation.

    The NREL calculations are just the straight net-present-value, and does not include any tax incentives or carbon offset sales etc. But if you plug in 1.50/watt (the likely unsubsidized cost to large competent developers) and a 20% capacity factor then use a favorable financing rate as the discount rate you can get to a very profitable 6-7 cent electricity after peeling off some of the up-front cost with tax incentives. Try it- assume $1500/kw, 25 years, $25/kw O & M, and a 3% discount rate, and the NREL calc gives you 6.5 cents/kwh LCOE. Whack 30% off the $/kw cost and the LCOE drops to 5 cents. From there raise the discount rate to 5% and you're still under 6 cents. As the industry continues to mushroom, this is what the LCOE of grid-tied rooftop solar will become before 2030.

    If you want real cost numbers on real projects, the greentechmedia solar wonks are all over it, and publish quarterly updates (and yes, the know how to do math.) Banking sector analysts are also pretty up-to date- if a bit conservative on their cost estimates. They're much closer to reality on both costs and projections than the plodding EIA folks.

Log in or create an account to post an answer.


Recent Questions and Replies

  • |
  • |
  • |
  • |