HVAC design engineers have used pressurized glycol systems for decades to transfer heat from boilers to the load. This method is the standard for hydronic heating systems and many industrial heat transfer tasks. Naturally, they applied this design to solar systems, treating the collectors as the boiler and the tank as the receiver. If you walk into most professional engineering firms and ask them to design you a hydronic heating system OR a solar system, they will grab their manuals that show how to assemble a pressurized glycol loop.
Since it is too expensive to fill the solar storage tank with a glycol solution, the heat is transferred from the collector loop into the tank through a heat exchanger. The collector fluid goes through one side of the exchanger and the tank water goes through the other side. There are two pumps, one on each side of the exchanger, and controls to turn the pumps on. Glycol loops are "charged" all the time. This is good. They remain ready to run whenever the pump kicks in. When such a system is installed, coin vents (can turn the screw with a dime) are installed at all the high points where air can accumulate and vapor lock the system. The startup procedure is to fill and pressurize the lines (maybe 15 psi) and go around to all the coin vents and burp the air out. Over the years, people have invented clever coin vents that when dry will leak air and when wet will seal. That way you don't have to go to each one to burp it, it will do so by itself. It is like the rope caulking used in boat hulls for thousands of years. As long as the boat stays in the water, all is fine. If you take it out and let the caulking dry out, it will leak until the caulking gets soaked again. There are many other kinds of automatic air vents, some based on the float system seen in toilets. Safety also requires a pop-off valve near the boiler (i.e. collectors) to relieve pressure in case the boiler controls go haywire. A glycol-water mix is a great solvent for shingles and plastics, including tile floors. So the pop-off valves must have a pipe running to a drain to contain the liquid in the event of a failure.
Since pressure goes up and down with temperature, a clever system was devised to maintain a nominal pressure in the loop. A tank, called an expansion tank, is installed in a tee in the line. The expansion tank has a rubber membrane running across the middle. The system fluid fills up one side and air fills the other side. The fluid in the system can expand and contract with temperature into the expansion tank, and the air bladder will keep the pressure within a specified range. The air pressure is set with an air hose and tire inflator, just like a car tire. A chart is used to determine the correct pressure according to the temperature of the system at the time. However, expansion tanks have a lifetime. The rubber (or neoprene, or whatever) bladder will someday crack from flexing as it ages and the expansion and pressure regulation benefits of the tank are lost. The system will usually vapor lock somewhere and the whole startup procedure has to be repeated.
Unfortunately, solar hot water systems don't like to play by the rules. They are not well-behaved. Typical HVAC glycol systems do not go through the extreme temperatures that solar collectors do. A boiler heating loop may have a maximum temperature of 140-160ºF. It never gets colder than room temperature inside a building, so the maximum temperature swing from summer to winter may be 90ºF (70-160ºF).
A solar hot water system, on the other hand, has the "boiler" sitting outside in the weather. It is always off at night where there is no sun. In the winter, the temperature may go down to -40ºF (Willmar, MN). Even in the mountains of NC, winter evening temperatures can go well below zero. A solar hot water system can have a maximum temperature swing as high as 260ºF (-40-220), or almost three times what a typical boiler system sees. In the summer time, the solar hot water system will see its maximum temperature, which varies according to the application. The most extreme case occurs when there is a very hot day with high solar radiation, and there is little need for the hot water. This can occur randomly on weekends, or summer vacations, and especially on space heating systems that sit idle all summer. When this scenario happens, the heat from the collectors is not needed and the temperature builds up until the boiling point is reached.
This same problem can occur if there is a power failure and the pump stops. At this point, a glycol system is in big trouble. If it gets to the boiling point, it will blow the pop-off valve. This drops the pressure in the system. The next night, there will be vacuum in the lines and the air vents will leak air in, vapor locking the system. The next day the hot glycol solution has air in it. A chemical reaction occurs with the oxygen that breaks the glycol into fatty acids, which can clog and eat the pipes if the situation is not corrected promptly. This scenario is not self-correcting. The system stops working, compounding the problem, and needs to be attended to. This is a progressive failure mode. The pump should never stop running during the day on a glycol system in warm weather. To avoid the over temperature problem, large glycol systems have additional equipment installed to dump excess heat. It usually consists of a big fan coil unit in the collector loop that kicks in when the temperature gets too high and dumps the heat to the outside world. The components include temperature controls, bypass valves, fans, and pumps. The added complexity just adds more failure modes. Heat dump systems cannot overcome power failures, unless you add a back up generator, which can have its own failure modes. At night in the winter when the collectors are cold, the cold glycol solution will try to circulate naturally down the supply line, creating a thermal convection loop. Some systems have even frozen the heat exchanger in this manner, causing rupture of the cold water line. A check valve must be installed in the collector supply line to prevent fluid from flowing backwards under cold conditions.
Whenever I think of solar glycol systems, I am reminded of the fairy tale about the little old lady who swallowed a fly.
The Development of Drain Back Systems
In an effort to overcome the many problems of glycol systems, early researchers turned to other methods. To overcome boiling and pressure problems with glycol, high temperature silicon oils were used. Unfortunately, they were very expensive, had poor heat transfer characteristics, and tended to leak out of soldered joints. Others tried air as the heat transfer medium. It won't boil or freeze. Blowers and ductwork to the collectors were a problem, and storing the heat from the air in a pile of rocks brought its own problems of mold and dust. You can't fab a rock pile and ship it to a site.
Others went back to plain water as the heat transfer fluid. It has the highest heat transfer capacity of any fluid. All others are measured against water, which is rated as 100%. Glycol is about 85%; silicon oil is about 20% as good as water. Since water will freeze and boil, the idea is to drain the water from the collectors at night, or when a high temperature limit is reached, so it is not there when the extreme conditions come. The system doesn't have to be pressurized, so tanks don't have to have an ASME pressure rating, which can double or triple the price.
Non-pressurized systems don't need pressure relief valves and expansion tanks. Early designs included air vents at the high points and heat exchangers between the collectors and storage. Some thought a vacuum breaker was required at the top to make the water drain out when the pump stopped. Some even installed a pipe between the collector supply and return lines with an electric valve to guide all the water to the return line for draining. All these vestiges of glycol systems only caused problems. Air vents and vacuum breakers introduce fresh oxygen into the water, accelerating corrosion. Conventional air vents on tanks cause evaporation losses, which required periodic refilling (and fresh oxygen). Protecting against corrosion by lining the tank is cost prohibitive above a certain size, and subject to cracking during transport. Check valves only complicate draining the water from the collectors.
The GRC Drain Back System
The way to make the drain back concept work was to rethink all the features of the system to minimize problems and maximize efficiency. This was the origin of the GRC drain back design. The original concepts were:
Non-pressurized operation with no expensive tanks, no code requirements, no pressure safety devices needed.
No heat exchangers between the tank and the collectors. Maximum heat delivered to the tank. Catch all the energy possible.
Maximum efficiency in delivery of heat to applications. Use no heat exchangers where possible, such as some space heating loops. A domestic hot water exchanger is always required.
Minimize evaporation losses from the non-pressurized tank. The tank vent design that emerged prevents ordinary evaporation losses while maintaining atmospheric pressure.
Simple corrosion control. A non toxic, food grade boiler corrosion chemical was selected that scavenges oxygen from the water, prevents galvanic corrosion, and helps clean the piping.
Unified tank system with multiple energy inputs and multiple outputs. This is referred to as "Grand Central Station", where all the energy is routed into and out of the tank storage system.
Simplest controls, no prioritization of energy output among applications. All applications have equal access to the energy. This prevents wasting stored energy by having one application holding off another.
Maximize thermal energy conservation. Enclose all pumps, exchangers, and controls within the thermal insulation of the system, where feasible. Use excess heat from pumps, for example, to heat the tank. Minimize line losses by including local plumbing inside the insulation shell. Some classes of pumps are water-cooled. They will work very well inside the insulating shell. Larger pumps are air-cooled and will not operate within the thermal shell of the tank.
The result is a system that is the simplest possible, the most economical to build, the highest efficiency, and the most durable. Many are still running after 25 years with only routine maintenance. In operation, when the collector pump turns off, all the water drains naturally back into the tank from both the supply and return lines. If there is a power failure, the water drains back in the same fashion as a normal shutdown. Neither heat dumps or antifreeze are needed to protect the system, since the water is not in the collectors when the pump turns off, for whatever reason. There is only one failure mode for a GRC drain back system. It occurs if the solar control activates when it should be off. Of course, this can't happen in a power failure. The "power on" failure mode has never been observed in the wild, but it is possible. A simple override control can be used to prevent the system from running when the tank is warmer than a certain temperature, or colder than a certain temperature. It simply interrupts power to the collector pump and requires human reset to start again.
This failure mode has been observed, however, due to operator error, so its effects are well known. In one case, someone left the control in the "Manual On" mode in cold weather. In another case, the factory controls were replaced by un-authorized heat pump controls in the field. If the "power on" failure occurs at night during freezing weather, and the system runs long enough to dump the tank heat and freeze the collectors, the result is a rupture of a collector pipe and dumping of the tank contents on the roof. When the tank contents are gone, the process stops. Since the water solution is non toxic, the run off is no more dangerous than rain water on the roof. If the control failure turned the system on during a hot day, the collectors might boil, sending steam out of the atmospheric vent, thereby preventing excessive pressures. If this condition lasted long enough, the tank water would be depleted and the steaming action would stop as in the freezing case. Current remote monitored controls are programmed to eliminate tampering and to report failures immediately. With only one obscure failure mode, and no dangerous results, the GRC drain back solar hot water system is far more "fail safe" than a glycol system under any climate or operating condition.
There are two elements to the GRC Drain Back design: the system concept and the resulting implementation into a family of products. The product consisted of a Fluid Handling System with all the components built in under the insulating shell. The electrical controls are mounted in a cabinet attached to the outside of the tank. A patent was granted for the drain back product design.
Dr. Ben Gravely
