Solar Hot Water

The Idea

We did some tests using a large (80 sq ft, 7.5 sq m) panel designed for heating swimming pools. It is relatively inexpensive ($140 US) and relatively large compared to panels traditionally used for heating domestic hot water. If it works, it will be used to heat water in our 880 gallon heat storage tank. A separate heat exchanger coil in that tank will preheat water before it enters our domestic hot water tank. The domestic hot water tank is has an internal heat exchanger itself, and is configured as an additional zone on the oil / wood hot water heating system.

Initial Blue Barrel Tests

The initial test was set up with a 37 gallon rain barrel on the deck about 8' above the top of the panel, which was placed directly on the ground on a south-facing slope. The panel was set up to thermosiphon with 5/8 garden hose. No real effort was made to optimize anything or reduce heat loss - the goal was simply to see whether the panels would generate enough heat and thermosiphon flow to be useful.

The pool heater has flow restrictors which must be painstakingly removed - it required an hour of work with hacksaw blades, chisels, and forceps to cut apart the restrictors and extract the pieces. Additionally, the heater has fewer holes of smaller diameter than desirable between the header pipes and the collector surface. Discussions will be had with the manufacturer if they ever return a phone call.

The results of the trial are in the database, and a graph for one day is shown below. Many lessons learned:

• Bubbles are a real problem.
• Standard garden hose gets really soft and kinks easily at 120°F.
• Even without optimizing anything, there is an impressive amount of heat output - over 3000 BTU/hour after losses
• Output temperatures of 120°F are easily attainable

Second Round Blue Barrel Tests

Thermistors were obtained and the system was instrumented to allow data collection. The graph on the left shows the results from a sunny day in May. Panel output temperatures exceeded 125°F, and the tank temperatures reached over 115°F, despite the complete lack of any insulation.

Solar Hot Water Installation, Phase 1

For this phase, there was only minimal instrumentation. There were three temperature sensors as shown in the adjacent block diagram. The sensor labled 'Domestic CW out' was connected to the outlet (top) of the preheat coil. The 'Domestic CW in' sensor was located at the inlet end of the coil. Both these sensors were mounted very close to the tank, so they give a reasonable approximation of tank top and bottom temperature as long as hot water is not being used. There was also a sensor attached to the tank itself at the midpoint.

Bifurcated Heat Exchanger

The solar panel inlet to the storage tank splits and connects to the heat exchanger at two points - the top and the middle. If the hot water coming in from the solar panel is hotter than the top section of the tank, it will rise and then circulate through the entire exchanger, top to bottom.

If, however, the water coming in is warmer than the bottom half of the tank but cooler than the top half, it will drop and circulate only through just the bottom half. This allows useful heat transfer for more hours per day or when solar conditions aren't ideal.

This graph shows a sunny day in August. Every time hot water is used, both the inlet and outlet temperatures drop. If you connect the high peaks, you see an approximation of tank top and bottom temperatures. Over the course of the day, the tank average temperature increased by more than 4 degrees.

Other observations: We could use more heat exchanger surface area for the domestic hot water preheat coil. During heavy use, water flows all the way through the coil and exits at a temperature of less than 80°F - about 15°F cooler than the tank top. The cold water coming in was less than 60°F, so even in heavy usage periods, we were getting almost 20°F of preheat. That works out to a heat transfer rate in the range of 20,000 BTU/hr.

Solar Hot Water Installation, Phase 2

These panels were obtained in used, salvaged, and leaking condition at no cost (thanks, Patricia) - there was a design defect that caused a stress concentration that broke the collector pipes inside the panels. A little work with the torch and milling machine and we're in business.

In addition to the glazed panels, other changes were made. A pair of mixing valves was added the previous winter. This eliminated the risk of scalding hot water when the storage tank is very hot. It also allows more efficient use of hot water in the DHW tank. More details on this can be found in the description of the domestic hot water system.

Cold Lock

During the previous year, we observed a problem that we dubbed 'cold lock'. As the graph on the left shows, temperature at the solar panel inlet to the storage tank drops dramatically during the early hours of the morning, and there is no heat input to the storage tank until the sun has been out for two or three hours. What's happening is this: The pipes between the solar panel and the tank are filled with water that has cooled to ground temperature during the night. In order to establish flow, the solar panels must get hot enough to force this slug of cold water uphill and through the tank, while forcing a slug of hot water that was in the tank heat exchanger downhill. In takes a long time for this to happen.

The solution is simple - install a section of pipe between the solar panel inlet and outlet right at the tank, but outside the tank insulation (visible in the block diagram above). This 'cold lock bypass' allows cold water to return to the solar panel without passing through the tank. As the solar panel heats up, this circulation will continue. As soon as the water coming from the panel is warmer than the tank (actually, warmer than the average of the bottom half of the tank), the water will circulate through the tank rather than the bypass - all completely automatically, with no moving parts and no controller. Thermodynamics at its finest.

Performance

With the glass panels, solar gain exceeded hot water usage for much of July and all of August, despite having large numbers of long-term guests - nine to eleven people in the house for much of that period. Panel output temperatures near 180°F were common, and the top of the 880 gallon storage tank exceeded 150°F. Many days saw 50,000 BTU of gain in the storage tank over and above hot water usage. For more than a month, 100% of hot water needs were supplied by the panels. For the solar season, oil consumption was reduced by 80% compared to pre-solar performance.

Results: Savings in fuel oil

The adjacent chart shows monthly oil burner cycles compared to the historical average for the solar hot water season of 2007 - approximately April 19th through October 19th.

The glazed panels were installed after the first month, on about May 19th. Software improvements were made in mid-June and again in mid-July.

In the 30 days prior to the glazed panel installation (April 19 through May 19), the oil boiler came on 47 times. This is an improvement over the pre-solar average of 65 times. The folowing 30 days saw only 27 oil boiler firings. At about June 19th, the first software change was made. From June 19 to July 19, the boiler came on only 9 times. The next 30 days, from July 19 through August 19, the oil boiler did not come on at all. During this period, the 880 gallon storage tank reached a temperature of over 150°F. As the sun angles became less favorable, the oil started to become necessary. The next 30 days, from August 19 through September 19, required 14 oil cycles. September 19 through October 13 (as I write this) required 16 cycles.

The five months after the glazed panel installation show an 80% reduction in fuel oil consumption for a savings of about 80 gallons. As the software improvements weren't in place for the first three months, next year should be even better.