Solar Hot Water Part VII: Retrospective

This post is about  "lessons learned" in the process of building our solar hot water system, discussing problems and some speculations about what might help to solve them.

As mentioned in the last post, a minor problem with using the system is that the gas-fired tank tends to cool off when the gas is turned off in the summer. In the long run, we plan to get rid of the gas-fired tank backup and put in a modulating on-demand electric hot water heater. My original plan was to couple the second heat exchanger in the tank with a geothermal heat pump or high efficiency gas boiler, but neither of these options is ideal. Geothermal heat pumps won't put out more than 140F and you need the heat transfer fluid substantially hotter to get good heating when the water flow is heavy, like when you are doing laundry and taking a shower at the same time. A gas boiler wouldn't allow the tank to derive optimal heating from the sun. If the thermostat is set at 120F (recommended for domestic hot water), then the gas boiler will come on when the water drops below that temperature and the solar collector won't come on. Same is true for solar tanks with a backup electric heater coil inside the tank. The solar tank can't preheat the water. An on-demand electric hot water heater should allow the solar tank to heat up as hot as the sun can get it, and just add enough energy to bring the domestic hot water up to 120F in the winter. In addition, we can add some solar PV to our roof to offset the electricity used to  run the on-demand heater. That should make our domestic hot water supply completely carbon neutral.

Aside from the aformentioned problem with an installer that did not know what he was doing, the primary problem with the system, though, is price. It is simply very expensive for the amount of carbon reduction achieved. Below is a graph comparing the cost per kilogram of eliminated carbon to two other alternatives: the same system but with a glass-lined tank instead of a stainless steel tank, and a cheaper thermosyphon system which a friend who lives in the same town had installed:

I used the estimated figure on gas eliminated from the Performance blog and included the federal tax rebate of 30% off installed system cost. I did not include the cost of the closet modifications for the thermosyphon system, because the storage tank is integrated into the collector so no change would have been needed in the closet. Note that the estimate here on cost per kilo carbon removed for our solar hot water system is too low, it does not include the winter when the solar thermal collector is contributing little.

Our system has a cost per kilo approximately that of the new refrigerator and I think it is for the same reason. I exchanged email with a friend who was surprised by the high price per kilo carbon reduction achieved for the refrigerator (which you can also see here). Normally replacing a fridge that is 10 or more years old by a new one is a fairly cost effective. I'm not sure how old the fridge was since it was there when we moved into the house, but on reflection, I think I know what the problem was. The new fridge was exactly the same model as the old, except updated with energy saving features. In our kitchen, the fridge is built-in, so we either had to get the same model or spend money on getting the space modified to take a different model. But the original fridge and now the new one are both high end models, and therefore very expensive.  It is a Lexus fridge, even though, we would have been happy with something a little less high end. Similarly, the solar hot water system is a high end system, and the cost reflects it.

That explains why the carbon reduction cost for our system is about three times as much as for the thermosyphon system, but it doesn't explain why even the thermosyphon system is expensive, at almost $20 per kilo carbon removed. At that price, even with the subsidy, it is not an affordable option for most homeowners. A high end gas-fired tank costs around $700 new, and a low end gas on-demand hot water heater costs about the same. An on-demand gas hot water heater can also result in substantial carbon reduction, though it won't be possible to completely eliminate carbon of course (it would for an electric on-demand heater though, by offsetting with solar PV). The thermosyphon system is around 10x the cost for an on-demand gas system, and our indirect solar system is more than 20x. Prior the WWII, many houses in California had solar hot water systems, and today, there are cities in China where solar provides hot water for over 90% of the homes. Why is solar hot water so expensive, the energy from the sun is free, it should be cheaper in the long run for a solar hot water system?

Well,  one problem is that natural gas (or, as I prefer to call it fossil gas) is just too cheap and likely to stay that way. Fossil gas releases only half the amount of carbon per therm of energy generated from coal, and something like a third less than oil so in some quarters it rates as a "clean" fuel. In addition, recent discoveries of gas deposits in shale across wide swaths of the Northeast and Mid Atlantic states promise that fossil gas will stay cheap for some time to come. On the one hand, that's good news because if fossil gas becomes cheaper than coal, switching electricity generation to fossil gas from  coal in states like Illinois where most of the electricity is generated by coal could result in substantial greenhouse gas reductions. On the other hand, unlike solar and wind, fossil gas is still putting fossil carbon  into the atmosphere. So to the extent that cheap fossil gas prolongs our dependence on energy technologies that dump fossil carbon into the atmosphere, it is doing us a disservice.

The other problem is that the technology itself is actually quite expensive and difficult to integrate into an existing building. It requires extensive plumbing on your roof and a backup system. Integrating with the most common type of backup system requires fussy yearly action to turn the pilot off and on, and results in suboptimal performance when the gas is off in the summer. Adding a backup system that works better costs more money. Compare this with grid tie solar PV, the kind most homeowners get these days, where the backup comes from the grid and so there is really no periodic maintenance needed or any kind of additional hardware. While solar PV requires structures on the roof (and, indeed, can result in roof leaks if not properly installed) and high voltage DC lines which are not trivial to get right, there is not as much dependence on the exact configuration of the existing structure except for proper solar exposure. A solar hot water system is more like the traditional off grid solar PV system, where the homeowner must install a large bank of batteries to store electricity for times when the sun isn't shining. Non-grid tie systems usually cost double or more what grid tie systems do, and the batteries also require monitoring and occasional replacement.

Since technology is what Silicon Valley is all about, you would think a budding entrepreneur would be studying this problem for the next great startup, but, as far as I can tell, there are no startups looking at how to reduce the cost on solar hot water. The economics seem pretty simple to me: get the cost down around where the payback time (sorry, unfortunately that's what most people look at) is around 3-5 years. Say, maybe the installed cost is 2x an on-demand gas heater, and you don't need a half inch gas main installed for it. Anybody up for the challenge?

Solar Hot Water Part VI: Performance

Last summer after the system came up, I turned off the gas hot water heater. The gas hot water heater has a pilot and pilots are just about the most energy wasting thing around. I was seeing temperatures on the solar tank of around 150F-160F in July, and it never rains in summer in California (like the old song says) so I figured we were good without the backup.


The results were mixed. Most of the time, we had plenty of hot water with the temperature being fine. With the shower on the top setting, it was not as hot as with the gas heater on (when you really need to turn the setting down to make it tolerable), but it was adequate for a nice hot shower after a long bike ride. Sometimes, though, the water was just lukewarm. This seemed to happen most often when we didn't use hot water for a number of days (like we showered at the gym) or when we went away for vacation. A couple times we had cloudy weather and the tank temperature dropped to 130F, which is I would have expected, but it didn't affect the domestic hot water temperature unless we didn't use much hot water.

The problem is that the gas-fired tank cools down if the hot water is not used for a couple days. Since there is a lot of water in the gas-fired tank, it takes a while for the hot water in the solar tank to draw through, even though the solar water is really hot. By and large, this wasn't much of a problem; it being summer, an occasional lukewarm shower wasn't an issue. I kept the gas-fired tank off until the end of October then relit it. At that point, the temperature in the solar tank was getting down below 130F.

Below you can see a comparison of our gas usage for July through November in 2008 and 2009:

As you can see, the usage was reduced from July through October, but it shot up in November. That's because we had a cold snap in the beginning of December (the November bill includes a couple weeks into December) this year, and November was generally much colder than last year. One of the difficulties of using year over year comparisons of energy usage is this kind of variability due to changes in the weather, etc. In fact, the increase in November completely overwhelmed the savings for July through October. July through November 2008 was 98 therms while the same period this year was 97 therms, just one therm less, or around 1%. If you look at the period from July through October, not counting the time when space heating was generating most of our gas demand, the reduction in gas usage is around 71%.

From last summer's data, I now know we use about 1 therm a month for cooking. In 2004, I discovered that our forced air furnace had a pilot light (yeech!) and I started turning it off in summer. I can therefore use the summer gas use from 2004 through 2009 prior to installation of the solar hot water system to estimate how much  gas we used for hot water heating. That gas will be completely eliminated in the summer 6 months (essentially April through September) with solar hot water. For the other 6 months, estimating usage is more difficult, because the solar acts as a preheater for the gas heater. So the amount of gas eliminated won't be the same as for the summer; on the other hand, it also won't be zero. To be conservative, we can say that it is, in fact, zero; that is, that the gas tank is responsible for all the hot water heating.

The average gas use in the summer months just for hot water heating is 6.32 therms per month. The average annual monthly use for 2004-2008 is 36.5 therms per month. Figuring 6 months of 6.32 therms reduction  and 6 months of 0 therm reduction, the percent savings comes out to 8.65%. So, say, around 10% reduction since we were on the conservative side about the winter use (yes, I know this is a fudge factor but it is the best I can do with the data I have). Anyway, I think I will need some more data for at least another year before I can really say what the reduction is like. There are also some ways to collect data that require measuring the hot water heater output, which I may decide to collect in the end.

The other interesting issue is the effect of the various insulation treatments I've tried. Below is a graph showing the temperature in the morning and evening in July before I installed the radiant barrier insulation:

I didn't use any sophisticated monitoring device for recording these temperatures, I just wrote them down on paper. That's why there are so many holes in the recording, some days I forgot, others we were on vacation. In general the temperature is running around 160F

Below is a chart of the morning and evening temperatures with the tank completely covered in radiant barrier insulation:

The gap between morning and evening is slightly smaller, and the average temperature of the tank drops over the month. The dates here are getting into fall, when the days are shorter and there is less sun on the collector.

Below is a graph of the morning and evening temperatures after putting on the R-13 fiberglass batt blanket:

The temperatures now are a lot lower because the time period is into December when the sun is at the lowest point, but the gap between the morning and evening temperature has narrowed considerably.

Insulation was highly effective in reducing heat loss from the tank. The temperature drop in degrees F per hour was 0.68 for no insulation - noticeably *above* what Superstor advertises. For radiant barrier only, the drop was 0.54, average of two sets of measurements, one shown above and one taken just before the batt was installed in December. For radiant barrier plus batt, the drop was 0.23. The temperature outside  the tank had little impact, the measurements in September and December for the radiant barrier insulation were close (0.57 in September and 0.52 in December). This might possibly be due to the fact that the tank closet itself has 6" of closed cell foam on the top and around 3-6" on the walls, keeping the tank enclosure toasty.

In fact, the insulation was so effective in cutting heat loss from the tank that I am now a bit worried it might overheat in the summer and cook the glycol. We shall see. I've signed up for a maintenance contract to have them come and check the glycol once a year to make sure it isn't acidic, at least for a couple years until I feel comfortable with the system. And, as a practical matter, the Schueco system is specifically designed to purge if the collector reaches stagnation temperature, to avoid cooking the glycol. So I think the system should be OK, and the extra insulation will sure help in the winter. Why can't Superstor put that extra insulation on the tank in the first place? It can't really be a question of cost. And I wonder, since it has about cut the heat loss in half, whether it would have allowed me to get by with one instead of two Schueco panels? Something for a pleasant evening of calculation in the near future.