1、2011 ASHRAE 657ABSTRACTResidential water heating is one of the least efficientenergy uses in residences in the United States. To gain a betterunderstanding of in situ performance, 24 natural gas waterheaters (eight storage water heaters of a single, popular modeland 16 instantaneous water heaters of
2、 nine different models)were installed in ten Minnesota homes. The water heaters wereextensively monitored and tested under an alternating modetest procedure for 15 months. The nine noncondensing instan-taneous water heaters used 22% to 54% less energy than thestorage water heater, saving an average
3、of 57 therms/yr(6.0 GJ/yr). The seven condensing instantaneous water heat-ers used 28% to 63% less energy than the storage water heater,saving an average of 74 therms/yr (7.8 GJ/yr). Even thoughinstantaneous water heaters can substantially reduce waterheating energy use, their high install cost is a
4、 major hurdle atcurrent energy prices. Measured annual water heater efficien-cies were also compared to the U.S. Department of Energy(DOE) energy factor (EF). Measured annual efficiencies ofboth types of instantaneous water heater were about 10%lower than their EFs, while those of the storage water
5、heaterwere 19% lower than its EF, suggesting that the relative perfor-mance of these two types of heater is not well characterized bythe EF metric. INTRODUCTIONWater heating is the second largest end use of natural gasin homes in the United States, accounting for 24% of residen-tial use (D water hea
6、ter has a 40 gal (150 L) tank.2. No min. start flow rate; water heater has a 0.5 gal (2 L) tank.Table 2. Important Instrumentation SpecificationsInstrument Parameter Measured Resolution Precision RangeWater meter (positive displacement)Total and water heater water volumes198.4 pulses/gal (53.4 pulse
7、s/L)2% of Reading0.5 to 25 gpm (2 to 95 Lpm)Gas meterTotal and water heater natural gas volume40 pulses/ft3(1413 pulses/m3)0.3% of Reading0 to 250 cfm (0 to 7 cmm)Watt transducer Water heater electrical consumption 0.2% of Reading 0 to 500 WImmersion RTDsWater heater inlet and outlet temperatures1/1
8、0 DIN148F to 752F (100C to 400C)660 ASHRAE Transactionsthe end of the previous monitoring period for the same heater.The StWHs had dials to set the temperature. These dials hadto be turned to a special setting to light the pilot at the start ofa monitoring period and were then returned as closely as
9、 possi-ble to the dial setting from the prior period. Homeowners wereasked to track any changes they made to the temperaturesettings on a log attached to each water heater.Residents at each test site were asked to complete a surveyafter the final monitoring period for each water heater. Thesurveys c
10、onsisted of two sections. The first section addressedthe acceptability of six aspects of hot-water performance. Thesecond section addressed the residents likelihood of purchas-ing or not purchasing the water heater given the heatersperformance on each of these attributes. The six performanceFigure 1
11、 Installation diagram.Table 3. Water Heater Installations by SiteSite Household Size, No. of People StWH IWH CIWH1 3 StWH1 IWH2 CIWH12 4 StWH1 CIWH33 3 IWH14 2 StWH1 IWH4 CIWH25 1 StWH1 IWH16 5 StWH1 IWH5 CIWH37 2 IWH2 CIWH18 4 StWH1 IWH5 CIWH49 1 StWH1 IWH410 2 StWH1 IWH3 CIWH4*The second IWH3 and
12、CIWH2 were never installed because of issues with ordering venting and availability of the WH.2011 ASHRAE 661attributes addressed were: delay time until hot water arrives ata fixture, the need to increase flow for low flow draws toreceive hot water, the consistency of water temperature forsingle dra
13、ws, the amount of hot water produced before runningout, the consistency of water temperature for multiple simul-taneous draws, and any reduction in flow rate for multiplesimultaneous hot-water draws. Data Analysis Annual natural gas use and efficiency were estimatedthrough a two-step process. In the
14、 first step the relationshipbetween daily output and daily input was determined, and inthe second step this relationship was used, together with aver-age output data taken over the course of the 15-month study,to determine annual energy use and efficiency. The sameprocedure was used to determine ele
15、ctricity consumptionrelated to hot-water output. Electricity consumption due tofreeze protection was treated separately. Analysis of the data verified a linear relationship betweendaily natural gas input and daily hot-water energy output (seeFigure 2 for an example) for the range of data collected.
16、Thisis consistent with results reported previously for commercialboilers (Hewett 2005), commercial water heaters (Bohac et al.1991), residential dual integrated appliances (Butcher et al.2006), and commercial gas cooking equipment (Horton andCaron 1994). These linear input-output relationships withn
17、on-zero intercepts produce typical efficiency curves whenplotted in the form of efficiency vs. output (see Figure 10 foran example). The slope of the input-output line can be thoughtof as something similar to the inverse of recovery efficiency,and the y-intercept can be thought of as the energy inpu
18、trequired to offset standby losses. For StWHs, the y-interceptclosely approximates the energy input required to keep theheater warm on a day with no draws. Except for the unit withthe buffer tank, the IWHs and CIWHs do not operate to keepthemselves warm, so the y-intercept does not predict energycon
19、sumption for a day with no draws, which would, in fact, beclose to zero. Rather, it reflects the energy required to make upfor typical daily transient losses. These transient losses occurdue to heating and cooling of the IWHs or CIWHs thermalmass when cycling in response to draws. On a daily input-o
20、utput plot, the y-intercept for IWHs and CIWHs thusaccounts for the fact that there are a certain number of hoursduring the day when the heater is cold and the losses are closeto zero, and daily data at low loads therefore fit the curve veryclosely except for those few days that actually had zero ou
21、tput. Input-output plots for individual draws (as opposed todaily averages) do show some nonlinear effects for very smalloutput energies, due to the thermal mass of the water heater,but over the course of a day the transient losses are either rela-tively constant from day to day or perhaps vary line
22、arly withload, so that the daily input-output plots do not show non-linearities other than the discontinuity right at 0 output.Thus, the linear input-output relationship accuratelycaptures water heating energy use for any practical daily drawvolume when the homeowner is in residence, but does notacc
23、urately capture energy use for days when the homeowneris away. The input-output lines for StWHs have more scatter (r2= 0.91 to 0.98) than those for IWHs and CIWHs (r2 = 0.97to 1.00). This is thought to be due in part to the varyingamounts of energy stored in the tank at the end of each day.Another f
24、actor is that the StWH setpoint is set by a dial and itwas not possible for the technician to return the dial toprecisely the same point after relighting the pilot for the startof each StWH test period. Different setpoint temperaturescorrespond to different standby losses and, therefore, differenty-
25、intercepts. For all water heaters, analysis of subsets of thedata and other side experiments showed that different inletwater temperatures, water heater set temperatures and ambient(a) (b)Figure 2 Daily natural gas consumption for three water heaters at Site #8; (a) I-P, (b) SI.662 ASHRAE Transactio
26、nsair temperatures created different linear relationships.However, the change in these linear relationships was smallenough over the ranges observed in the ten sites monitored thatvery high r-squared values were computed for all regressions. Because the relationship between input and output islinear
27、, the mean energy use for any period can be computeddirectly from the mean heater output. Daily hot-water energyoutput varied linearly with the temperature of the cold watercoming into the house (referred to here as the “main temper-ature”). This relationship has been shown previously inMinnesota (H
28、ancock et al. 1996). The main temperaturevaried with season and depended on the water source. Eightsites were supplied by city water from surface sources. Themain temperature for these homes ranged from 37F to 72F.One site had city water from a municipal well. Its maintemperature ranged from 47F to
29、57F. Another site relied ona private well and its main temperature ranged from 47F to52F. The two homes with well water sources did not have alarge enough variation in main temperature to produce a statis-tically significant correlation between hot-water energyoutput and main temperature. The main i
30、nlet temperature tothese homes did not vary enough over the course of the year toaffect that water usage in the homes.Day-to-day hot-water energy output varied considerablydue to variations in water-use activities in the home. Weeklyaverage output was much less variable and so better suited toanalys
31、is of seasonal variations. In order to determinewhether hot-water energy output was statistically differentfor different heaters at the same house, weekly output wasregressed on main temperature with the water heater used asa dummy variable. Because of the linear relationship between the hot-wateren
32、ergy output and the main temperature, the mean output canbe computed directly from the mean main temperature. Thelinear relationship between hot water energy output and maintemperature may reflect both greater energy input required toheat colder water to a given setpoint and increased hot-watervolum
33、e used due to the need to blend more hot water with thecolder cold water for a given shower or bath water tempera-ture, or perhaps taking warmer showers or baths in colderweather. If the water heater dummy variable was not signifi-cant, the same mean output was used for each heater. If thedummy vari
34、able was significant, the water-heater-specificmean output was used. These output values were then usedwith the linear input-output relationships for each heater tocompute mean annual energy use. The main temperature was a measurement of the watertemperature coming into the home. Main temperatures w
35、erecalculated by looking at all draws in a day that were over threeminutes. It was assumed that after the first minute of a draw,all water that had been in the pipes between the inlet of thehome and the water heater, which had been warmed by theambient room conditions, had passed through the water h
36、eater.This allowed an average main temperature to be calculatedfrom the average temperature at the water heater inlet over theremaining duration of the draw. These long-draw temperatureswere averaged over each day and defined as the daily maintemperature. Mean annual main temperatures were computedf
37、or each site by averaging the daily main temperatures over thecourse of a year. In cold northern climates, such as Minnesota, an IWH orCIWH heat exchanger could be damaged if standing waterwere allowed to freeze inside the unit. Under some venting andusage scenarios, cold air can enter the units thr
38、ough thecombustion air supply or exhaust. If temperature sensorsinside the water heater drop below a manufacturer-determinedlevel, electric heaters inside the unit are triggered. The runtimeof the electric heaters and the power draw required duringfreeze protection vary from heater to heater. A refe
39、rencetemperature for each IWH was defined as the daily averageoutdoor temperature below which electrical consumption forfreeze protection was observed. Electricity consumption forfreeze protection increased linearly with decreasing averagedaily outdoor temperature below the reference temperature.Any
40、 electrical consumption for freeze protection was includedin the energy use and savings calculations for IWHs andCIWHs.Error EstimationHigh resolution instrumentation was selected to reducemonitoring errors from field data. Data were also collectedover 15 months to increase the total days of data. T
41、he largenumber of monitoring days allowed for high coefficients ofdetermination (R2) for regressions. These two factors allowedfor a maximum daily input and hot-water output demand errorof two percent.RESULTSWater Heater Energy Use and IWH/CIWH SavingsA sample plot of daily natural gas consumption a
42、s a func-tion of daily output for all water heaters at one site is shown inFigure 2. The demand-related (non-freeze protection) electric-ity consumption for these same heaters is shown in Figure 3.Demand-related electrical consumption accounted for 44% to100% of the total water heater electrical con
43、sumption for eachof the test sites. The natural gas and electricity consumptionrelations were used with annual output data to determineannual energy consumption.At nine of the ten sites the hot-water energy output as afunction of main temperature was the same for all water heat-ers. For these sites,
44、 output data from all the water heaters weregrouped together and regressed on main temperature. Asample plot of seasonal hot-water energy output vs. maintemperature for this type of site is shown in Figure 4. At onesite there was a statistically significant difference in output fordifferent water he
45、aters (Figure 5). At this site the water heaterswere treated separately when determining annual output fromannual average main temperature. 2011 ASHRAE 663For each site, mean main temperatures were computed,then the average weekly hot-water energy output correspond-ing to this main temperature was d
46、etermined (for all heaterstogether or each heater separately, as described above), andfinally the average daily natural gas and demand-related elec-tricity consumption for each heater were computed for eachheater from the average daily hot-water energy output. Aver-age daily natural gas and electric
47、ity use were then comparedat each site to determine the savings for IWHs and CIWHs. InTable 4, average daily parameters are shown for each site.Table 5 shows the annual natural gas and demand-related elec-tric consumption for each water heater at each site and thesavings for IWHs and CIWHs relative
48、to the StWH. Table 4does not include any freeze protection energy consumption.The usage from freeze protection is handled separately. Atsites 3 and 7 where no StWH was installed, an average input-output relationship from the eight StWH sites was usedtogether with the hot-water energy output at sites
49、 3 and 7 tocompute the estimated StWH energy use and correspondingIWH and CIWH savings. All savings calculations are siteenergy savings and do not account for site to source energyratios.Electricity consumption for freeze protection wasobserved for one CIWH and four IWHs. Water heaters thatrequired freeze protection had a significant increase in elec-trical consumption when the average outdoor temperaturedropped below freezing. Figure 6 shows the relationshipfound between electrical consumption and outdoor tempera-ture for the IWH at Site 6, where the freeze protection wasneve
