«May 2016 NOTICE This report was prepared as an account of work sponsored by an agency of the United States government. Neither the United States ...»
3.3 Comparison of Set Point Temperatures The tenants of LaFayette had no access to the HPWHs, which were located in locked mechanical closets to protect the equipment; this is a common practice in rental housing. The homeowner of the NCTH in Savannah (E) had access to the HPWH in the attic and could adjust the set point temperature and operating mode. Data indicated the HPWH was unintentionally set to vacation mode on August 21, 2014, because several DHW draws occurred between August 21 and August 23 without the heat pump operating and the HPWH outlet water temperature decreased significantly (Figure 10). The homeowner was apparently not satisfied with the 82°F water at the spigot, so they turned the HPWH to efficiency mode and increased the set point temperature on August 23, 2014. The HPWH power consumption and output water temperature increased, but the patterns remained the same after the event, whereas a change to hybrid or electric mode would have caused a different power consumption pattern. The HPWH compressor’s power consumption was higher because more work was required to move heat into a tank at a higher temperature.
Figure 10. Changes in the HPWH operating mode and tank set point temperature can be seen in the raw data Unit E was segregated into two periods (E1 and E2) for analysis based on the differing tank set points (Table 7) and other critical daily averages that affected HPWH performance.
E1 reflects the period from July 9 to August 20, and E2 reflects the period from August 24 through October
15. Figure 11 shows a graph of the processed daily averages of HPWH inlet and outlet temperatures used for the daily COP calculations. The average water outlet temperature during E1 was 113.5°F when the tank set point temperature was 120°F. The lower measured temperature was due to thermal losses from the measuring equipment and short water draws, which are addressed later in the document. The average daily water outlet temperature during E2 was 133.1°F (17.3% increase), it peaked slightly higher than 140°F, and the set point temperature was verified to be 150°F (the HPWH maximum). The daily average change in water temperature across the tank increased by 19.4°F, a 54.4% increase; the inlet water temperature increased by only 0.4%. The higher set point is associated with a lower average daily COP value of 2.0 compared to the 120°F set point average daily COP of 3.1.
Figure 11. Processed data of daily HPWH incoming and outgoing water temperatures depicting a change in tank set point temperature.
E1 includes period to the left of the first dashed line. E2 includes period to the right of the second dashed line.
The average daily volume of hot water consumed decreased by 21.8 gal/day; the daily average DHW draw time decreased by only 2.1 min after the tank set point was set to 150°F. Assuming the average temperature at the tap was the same during both periods, an energy balance of mixing hot and cold water streams would require less DHW at the higher temperature.
Figure 12 displays the trend between daily COP values and daily water consumption for both set points. The COP increases as the consumption increases, and the COP values are generally higher for the lower tank set point temperature. Lower COP values at the higher set point for similar consumptions can be attributed to higher tank losses and a high refrigerant condensing temperature at higher tank set points.
Figure 12. Daily DHW use versus daily COP values for both set points at Savannah E
3.4 Combined Site Results Major variability in daily COP values was observed across the monitoring period (Figure 13 and Figure 14). On most occasions, significantly higher daily COP values were observed at times and lower COPs at others. This random distribution can be largely attributed to the variability in daily DHW use seen in Figure 12. Another reason is the timing of the HPWH turning on.
According to A.O. Smith documentation, the HPWH is set to turn on when the tank temperature reaches 9°F below set point. On many occasions the heat pump would operate at the beginning or end of a day; thus, the energy consumption was attributed to a different chronological day than the hot water draw actually occurred. This was largely dependent on the behavioral consumption of the hot water.
Figure 14. Bar plots of daily COP values of all four units in LaFayette.
The bare areas indicate periods that did not meet all criteria to be considered valid data.
When the daily COP calculations were averaged, equal weight was given to the daily COP values. The average daily COP values for a 120°F tank set point at the LaFayette site were 1.9 and 2.5 and 3.1 at the Savannah site. Thus, the COP across the entire monitored period was calculated and resulted in a slightly different COP values at each. These results and the daily average values of variables used in the COP calculation are listed in Table 8.
Figure 15 displays the trend between daily DHW consumption and average daily COP. The COP increased sharply before reaching the knee of the curve between 20–40 gal/day before leveling.
This trend was reported by Shapiro and Puttagunta (2013). They reported an average COP value of 2.1 for the unducted A.O. Smith Voltex models.
Figure 15. Scatter plot of daily DHW use versus COP for all five units
Figure 16 displays the dependence of daily HPWH energy consumption on daily DHW consumption and daily average T of tank inlet and outlet water. The energy consumption increases linearly as the DHW consumption increases. The energy consumption increases at a faster rate as the of tank inlet and outlet water increases. Royal blue dots (recorded from house D) in the lower left corner represent days when few water draws occurred and the heat pump never operated, when no draw occurred but the heat pump operated due to tank losses, or a combination of both.
Figure 16. Daily HPWH energy consumption as a function of daily DHW use and daily average across the tank Figure 17 plots the intake wet bulb temperatures directly against the daily COPs.
Unit E1 (pink), in Savannah, Georgia, with a 120°F set point, appears to have predominantly higher intake air wet bulb temperatures and daily COP values than the other units. E1 did have the highest daily average COP. Otherwise, no individual unit shows an apparent trend linking increased intake air wet bulb temperature to increased COP values.
Figure 17. Intake air wet bulb temperature and daily average COP
3.5 Effect of Duct Configurations in LaFayette All HPWHs at the beginning of the monitoring period were configured with an unducted intake and ducted exhaust directly to the encapsulated attic. The intake pulled air from the attic through a transfer grille in the ceiling. Configurations at Units A and B were adjusted on September 17, 2014, to evaluate the performance impact of various duct configurations. A duct kit was installed on the intake sides of both units and connected to the hole where the transfer grille was removed via 8-in. insulated (R-6) flex duct (Figure 18 and Figure 19). A.O. Smith said that the relief vents on the bottoms of the duct kits could be sealed off without harming the HPWH and provided flexible adhesive strips to seal the vents (Figure 20). The flexible strips proved difficult to install but successfully blocked all air leakage, so duct mask was applied. Both the supply and exhaust were ducted in this configuration for Units A and B.
On September 26, 2014, the exhaust duct on unit B was removed so the HPWH exhausted air directly into the mechanical closet. The duct that connected the closet to the attic remained open to determine the impact of increased pressure from excessive ducting.
Figure 19. Intake duct configuration at Site B.
Transfer grille in the ceiling before being removed (left); ducted intake grille and sealed exhaust relief vent (right)
To compare HPWH performance under various ducting configurations, variables were analyzed for equivalent durations directly before and after the duct configuration changes. Table 9 documents the date ranges of the HPWH duct configurations and the corresponding average daily COP values. The COP values remained the same for Units A and B with intake ducting and for Unit C, whose duct configuration remained unchanged. A completely unducted HPWH was not tested, because it went against the collaborator and manufacturer’s recommendation due to the low volumetric air capacity of the mechanical closet.
Figure 21 shows box plots of the intake air temperatures of Units A and B over the preceding period and when the heat pump air intakes were ducted directly to the encapsulated attic. The mean intake air temperature for Unit A increased from 68.9°F to 70°F when the intake was ducted. The mean intake air temperature for Unit B stayed at 69.7°F during both duct configurations, but the daily deviations were much greater. The average air intake of Unit C, which remained unducted during both periods, decreased from 72.8°F to 71.4°F. Aside from general increases and decreases, Unit C cannot be compared further to Units A and B because its intake temperature was always greater, and Units A and B were in the same duplex and shared an uninsulated interior attic wall. Unit C’s intake air temperature was warmer over the entire monitoring period even when the other units were ducted. The average daily intake temperature likely had an insignificant impact on the COP because the temperature change was minimal.
Ducting the intake would have increased air temperature across the evaporator coil because the recirculating exhaust was eliminated, but the average attic air temperatures varied up to 6°F/day and the intake temperature depended on the time of day the heat pump operated. The increased static pressure across the fan—if it was measurable—was caused by the intake duct and did not affect the HPWH’s efficiency.
Figure 21. Intake air temperature of Units A, B, and C over the periods T1 = 8/26/14–9/16/14 and T2 = 9/18/14–10/19/14.
Units A and B had ducted intakes during T2 and Unit C’s intake remained unducted for the entire period (+ indicate outliers).
3.6 Impact on Encapsulated Attic Air Temperature and Humidity The impact of the HPWH on the temperature and humidity of the encapsulated attic and the living space was evaluated. Dehumidification in Unit E’s attic is clearly observed when the HPWH operated during the summer (Figure 22); however, the airstreams returned to their original states quickly after the HPWH stopped operating. This indicates moisture removal was rather insignificant relative to total attic moisture content. Figure 22 also depicts the temperatures and absolute humidities of the air that entered and exited the HPWH. The exhaust air sensor was installed inside the duct, and its absolute humidity spiked after HPWH operation. During heat pump operation, latent heat was removed from the airstream via condensation that drips from the heat exchanger coils to the drain pan. Internal factory-installed drain lines did not meet flush to the pan to allow it to drain completely. Further, condensation remained on the heat exchanger coils when the heat pump stopped operating. This confined area also houses the compressor, which reached temperatures near 200°F (Figure 9). At the conclusion of a heat pump cycle, the condensed water in the pan and on the coil re-evaporated from the heat of the compressor, became trapped in the now-cooled duct, and caused the sensor in the exhaust duct to measure high absolute humidity. Condensation in the cool duct was detected and mold growth within the duct is possible, although none was observed.
Figure 22. Savannah E HPWH intake and exhaust temperatures and humidities
Figure 23 compares the absolute humidities at the high center location of the attic and of the living space of Unit E (Challenge; the home was seeking U.S. Department of Energy Challenge Home certification) and of the neighboring home (Unit F). The neighboring home is of similar construction (same size, HVAC equipment, and sealed attic) except it has an ERSWH in the attic. The peak absolute humidity level in the attic was reduced only when the HPWH operated before peak times; it generally needed to run during the morning when the humidity levels were rising. When the HPWH operated at other times during the day, the absolute humidity levels almost immediately returned to the same levels as in the neighboring home. The HPWH did not provide significant dehumidification; the humidity decreased only slightly when the HPWH was running. The humidity levels in the living zones were entirely dependent upon the latent heat removal by the HVAC systems.
The fluctuations of the absolute humidity levels in the attics are believed to be highly influenced by the “sponge” effect of open-cell foamed rooflines, which is due to the moisture loads being driven in and out of the foam by radiant solar heating and night cooling. The moisture levels in the sealed attics at all six sites showed similar daily moisture levels throughout the year. Under the current regime of operation the HPWH does not appear to remove enough moisture each day to make significant reductions in daily peak moisture loads compared to the adjacent house with a standard ERSWH. Further monitoring and research are needed to better understand this effect and to develop strategies that effectively reduce moisture levels in sealed attics of low-load homes. One potential HPWH operating regime would be to significantly increase the tank set point temperature in the mornings.
Figure 23. Savannah Unit E and F absolute humidities at the high center location of the attic and of the living space The attic temperatures were monitored in four locations (north, south, and east sides of the house and high center of the attic about 6 ft from the attic floor, Figure 24).