Challenge

The original project goal was rooted in optimizing a current design flaw in residential hydronic heating systems. In a residential hydronic heating system, hot water is taken directly from the hot water heater, and is used as the working fluid. The fluid is run through an expensive, hydronic coil, designed to withstand the particle buildup and corrosion effects of water. The water is then recycled to the hot water heater.

This system works fairly well in the winter when the heated air is needed frequently to condition a home. During warmer months, however, the water in the system sits stagnant, and must be cycled occasionally to prevent bacteria build-up (namely legionella), counteracting the effects of the AC if it is running. The original project task was to design a method of flushing the system, to prevent bacterial growth, without counteracting the AC system.

Solution

Our idea generation step allowed the realization that isolation of the water loop to avoid contamination was likely the most effective solution. The original problem was solved fairly quickly using several solenoid valves and an Arduino board. When cycled, the design would activate the normally open solenoid to close off the water return route. Simultaneously, two normally closed valves would be activated, opening a drain route, and supplying the system with pressurized city water. This system would flush any contaminated water without it coming into contact with the hot water in the tank to be used in the house. The city water would be at a much lower temperature and would have a chilling effect, if any, on the air as it moved through the hydronic coil.

Upon further idea generation, we decided that if the loop would be isolated, it no longer needed to be water, but could be any fluid. Refrigerants have a much more beneficial thermal conductivity, so the idea of a refrigerant fluid loop began to develop. After further consideration, we found a design that would act like a heat pipe in essence, however, instead of using wicking material, gravity would provide the necessary pressure head required to return the fluid to the heating end. This simplified the design drastically, as it removed the need for a compressor or a pump, and leaves an entirely passive system with very little room for mechanical failure. The design is intended to be used on a gas water heater, which rejects flue gas in excess of 300 degrees F to the outside atmosphere. By using that excess heat to boil the refrigerant, no additional energy is required to heat a home when the hot water heater is already running, which would result in massive energy savings. The test setup is pictured, and uses an electric water heater with the top removed via plasma cutter to provide a proof of concept.

Result

The project concept was a success. While the design initially contained a check valve, this added complications to the flow. The check valve was removed, and the system was heated with both the inlet and exit pipes completely open. After a short amount of time, the inlet and outlet temperature difference reached a maximum of 55F, with the inlet side at around 127F and the outlet temperature at 72F. The only way this temperature difference is possible with both sides completely different is if the heated refrigerant was successfully flowing through the inlet port, and out of the outlet port, in a saturated liquid form. This can only happen if the heat is removed from the refrigerant, which means the heat must have been moved to the air.

Unfortunately, we were not able to see as promising results with regards to the air temperature rise. While we did see a 8-9F temperature prior to the official data collection, we were only able to obtain a 2-3F rise in temperature during the official testing. We believe this insignificant amount of temperature rise to the air is directly related to the condensor coil we were using. The condensor coil was donated to save cost, and ultimately was of poor design for our specific purpose. It was a two-track design, with far too many turns. The volume of the inside of the condensor coil was around the same volume of the rest of the tubing coil within the water heater. Also, the bottom track was capped off to maintain coil pressure. Because the tracks are interconnected, we strongly believe that the majority of the refrigerant that the system was charged with was condensed in the coil, and "stuck" in the capped off track.

We understood from the beginning of the project that mass flow rate of the refrigerant was directly related to the total heat transfer available. The refrigerant is the medium of heat transfer, so the more heat the refrigerant could move from the water heater to the condensor coil, the more heat that can be used to raise the temperature of the heat. With the majority of our refirigerant mass "stuck" within the coil, our mass flow rate was significantly diminished. We have evidience of this suspecion, based on the second video above, with the cyclic flowing of the refrigerant passed the sight glass. We believe that there was a very small amount of refrigerant available to circulate, and we can actually see it for the short time that it is moving passed the sight glass. Further evidence is the observation that as the small "surge" within the sight glass passes the outlet thermocouple, the temperature rate slightly drops for a moment. The timing of the surges was very predictable.

If the system were to have the correctly designed coil, we believe all the condensed refrigerant would successfully drain back to the water heater for recycling. We charged the system with a little more than 5 lbs of R22 refrigerant. Much like a cup needs to be entirely full before it allows any water to overflow, the condensor coil track continued to hold enough refrigerant until the track was full, and any excess could then be used for circulation. If we were to receive steady flow of the system, rather than cyclic flow, a much higher mass flow rate would be available to move heat from the water heater to the air.

Because the project was a proof of concept, the concept was successfully proven. It is entirely possible to saturate refrigerant with heat, passively allow the refrigerant to flow to the coil, condense, and return to the water heater, effectively transferring heat to the air without any external power beyond heat removed from the water heater. The project will be passed onto another senior design group for future work, and a list of recommendations can be found in the report above.