Hot Water Heater
While we’re talking about the efficiency of heating water, how well does my gas-fired hot water heater perform in transferring combustion energy into the water tank? Seems like it should be pretty good: few places for the heat to go except into the water (although the flue does get super-hot).
Practically speaking, efficiency suffers from slow heat loss during idle times, and from loss through the pipe walls during delivery. Tankless applications can avoid these two loss mechanisms. But let’s separate these out and ask only about the direct efficiency with which combustion energy gets into the water in the first place. To do this, I poured hot water into a bathtub until the water heater came on. I let the heater complete its business, then waited about an hour, after which I drew another batch of hot water until the heater activated again. This way, the second heating did not have to compensate for long-term loss of heat from the reservoir.
Mind you, this experiment represented unusually excessive hot water use in our domicile, but I did it for the people. And we did manage to enjoy baths in the bargain.
I measured the temperature of the water emerging from the (hot only) tap, and from a tap right at the inlet from the street (adjacent to water heater, too). I used the water meter by the street to gauge volumetric use to a precision of 0.01 cubic feet (0.3 ℓ). And, of course, the gas meter told me how much energy was consumed.
On the first wave, I used 1.39 ft³, or 39 ℓ, heated from 22°C to 53°C (remind me to turn the heat down; I should also note that the water came out at 50°C on the first draw, but it had cooled in the tank for some unknown time). Multiplying these fine numbers by 4184 J/ℓ/K, I compute an energy demand of 5.1 MJ. Meanwhile, my gas gauge made 17.15 revolutions at 510 Btu/rev for a total of 9.2 MJ. The efficiency computes to 55%.
On the second withdrawal—this time based on a recently heated tank, it took 51 ℓ before the heater engaged, after which the gas dial made 19.15 turns. This time, the demand was 6.6 MJ, while the gas cranked out 10.3 MJ. The efficiency is 64%.
Again, I find myself disappointed. I need to re-adjust my intuition about how straightforward it is to channel heat from a flame into water on the other side of a metal wall.
Bonus Round: Gas Oven
I seriously doubt I’ll get around to another post detailing the things I learned from my laser-gauge gas meter. So I’ll stick in here what I learned about my gas oven. How much energy does it take to “charge” up? How much power to keep steady? How much “on” time equivalent does it take for the pre-heat phase?
I heated my oven to 425°F (218°C) from an ambient 20°C temperature. It took 9.5 minutes to arrive at the setpoint, during which time the half-cf dial on the gas meter made five turns. Correcting for the water heater pilot light rate, this action took 2480 Btu, or 2.6 MJ (0.72 kWh). The burner operated at about 4800 W (16,300 Btu/hr). Thereafter, I found that it took a steady 1500 W to maintain temperature: I left the oven unopened and undisturbed for the better part of an hour to make sure I reached equilibrium—all for the sake of experiment. This means that the preheat phase uses the same amount of energy as 30 minutes of steady operation. You don’t need fancy gauges to tell you this if you note the preheat time is 10 minutes and observe that the burner is on one third of the time during steady operation.
If you’re heating a pizza in the oven that requires ten minutes of cook time, then only 25% of the total energy is spent in cooking mode, the other 75% in preheat. If cooking something for an hour, the preheat surcharge drops to 15%.
Although I am embarrassed to reveal the efficiency of cooking pizza in the oven (since this is not unknown in my household), I owe it to myself to carry out the calculation. Let’s say I heat a 383 g pizza by 200°C (but that it’s heat capacity is in between that of water (at the high end) and more typical materials—say 2000 J/kg/K. So I need to inject 0.383·2000·200 = 153 kJ. Meanwhile, my oven heats up for ten minutes and then the pizza spends ten minutes in the oven. I count 2.6 MJ for preheat, and an additional 0.9 MJ for the cook time. In the end, I manage to get 4.4% of the expended energy into the pizza. If I instead tried 6 minutes of full-power equivalent in the microwave oven (at 1750 W), I might get near 25% efficiency—and a flaccid crust.
Lots of numbers thrown around in this post. Here’s a table of my results.
|Boiling water on gas stove, full blast, no lid||16%||—|
|Boiling water in same pot, smaller burner, with lid||27%||—|
|Boiling water in kettle on small gas burner||27%||—|
|Heating water in microwave oven||43%||15%|
|Boiling water in electric kettle||50–80%||18–28%|
|Hot water heater, including tank loss||55%||—|
|Hot water heater, without tank loss||64%||—|
The adjusted efficiency is the fossil fuel equivalent if electricity is derived from fossil resources (coal, natural gas) at 35% efficiency. The range on the electric kettle depends on how quickly the kettle shuts off in response to boiling water.
Testing Without Lasers
If you wanted to replicate or extend these experiments, is it hopeless without the laser-sensor gauge I created? Not at all—although slightly less convenient. In fact, I carried out the hot water heater test after I had already dismantled the gauge. Here’s one trick. Once you’ve characterized the burn rate of various devices (e.g., stove burners on max flame; oven while burner is on; water heater; furnace; etc.) then all you need is a way to measure time when the (audible) burner is on.
As alluded to above, keeping track of the fraction of time the oven burner is on in steady operation is enough to tell you how much energy goes into preheat vs. holding temperature. Even without calibrating a stove burner, different configurations (lid, different pots, etc.) can be compared against each other just by timing.
To measure the rate of gas usage of various devices, make sure that device is the only thing on, and time how long it takes the half-cubic-foot (or 2 cf) dial to make one revolution. Together with knowledge that each half-cf translates to 510 Btu (538 kJ), you’re pretty much set. Having a way to measure volumes and temperatures also came in handy for me.
What of It?
Heating water is less efficient than I originally thought. All the same, it will always take 1 kcal to heat 1 kg of water 1°C, and heating water is something we will always be interested in doing. Yet, efficiencies are middling-enough that we can’t expect gigantic improvements. Heat pumps could break the 100% efficiency barrier (by a factor of several), but these are impractical for small-scale applications.
Should you react to the numbers above by rushing out to buy an electric kettle, with the promise of tripling the efficiency over the stove-top solution? In part, this depends on your source of electricity. If your electricity is derived from fossil fuels, and you otherwise have a gas stove, then it’s probably not worth it. But even if going from an electric stove to en electric kettle, we must consider the embodied energy of the kettle. I discussed two methods for estimating embodied energy in another post, which for this case results in something like 50 kWh of investment. Each cup (250 mℓ) of water takes 0.03 kWh of energy at 70% efficiency. If you’re tripling efficiency, then you save 0.06 kWh per cup, and need to boil over 800 cups before the thing pays for itself, energetically. Maybe this makes sense. But it’s not a burning imperative.
As with many things, a far more effective strategy is to first recognize how and why you use sources of energy. After developing an awareness, you are far less likely to heat excess water in a kettle that you don’t plan to consume. Shorter, less frequent showers can have a far bigger effect on your energy use than how you heat water for tea. It comes back to behaviors.
In the meantime, it’s kind-of nice to have some numbers for water heating efficiencies—as disappointing as the numbers themselves are.