Printer Friendly

Measured residential hot water end use.


Water heating is the second largest end use of natural gas in homes in the United States, accounting for 24% of residential use (D&R International 2006). Despite this, detailed residential hot water usage is not well understood in many homes.

In 1999 the American Water Works Association collected water end use data from almost 1200 homes in twelve U.S. cities. This study collected total water usage, and no designation was made between hot water and cold water. The study made significant progress into understanding total water use in homes. It included information about water leakage rates and fixture usage patterns and offered good recommendations for water conservation programs (Mayer and Deoreo 1999).

Flow trace signature analysis uses the flow rate, volume, and length of water uses to define a signature for each type of hot water use. One limitation of the flow trace method was that it cannot distinguish between similar fixtures. This method can determine that a draw was made at a sink but not if it was a bathroom sink or a kitchen sink. Flow trace has been a well-documented method for studying residential water use for many years. In 2005, flow trace was used to determine the savings of both water and energy from water conservation products. This EPA study collected hot water usage data from 20 west coast homes. The study monitored two weeks prior to installing conservation product and two weeks post retrofit. In additional to collecting some fixture usage data, this study showed a 21% reduction in hot water usage from water conservation products (Environmental Protection Agency 2005).

The flow trace analysis was improved by adding temperature sensors along the distribution system branches to further disaggregate (Lowenstein and Hiller 1998). This improved the methodology so that draws could be assigned to specific sinks, which furthered knowledge about the difference in sink usage. Fifteen-second hot water consumption data collected over 2.5 years from fourteen single-family residences and a fourteen-unit multi-family building allowed for the development of new hot water consumption data analysis methods. This study developed an improved system of water heater sizing, concluding that the first hour capacity of a water heater was generally not useful. The critical time period was typically less than 35 minutes (Hiller 1998).

In 2004, improvements to the flow trace methodology were taken a step further, and a new hot water disaggregation method was developed (Tiller et al. 2004). This methodology called for one-second interval data with temperature sensors for each hot water end use.

A better understanding of the typical number of hot water draws per day, daily hot water volume, draw length, volume, flow rate, and spacing would help manufacturers, energy program administrators, and policy makers design better products, programs, and policies to improve water heater performance and reduce energy use. Hot water usage and energy are closely tied. If hot water usage can be decreased, the energy needed to heat water will also be decreased. Wasting hot water wastes energy. A better understanding of how hot water is used can potentially save both water and energy.

Real world data are a necessity to understand hot water usage patterns. The frequency and length of showers and the prevalence of baths vary from household to household. Clothes washer and dishwasher draw patterns are complex and vary with the type of cycle selected. People often turn on the hot water faucet for short sink draws where they never actually receive any hot water. Many of the uses of hot water are so automatic that people barely realize hot water is being consumed. For all these reasons, monitoring is necessary to accurately characterize usage.

Incorrect assumptions about hot water usage may have important unintended consequences, and better data can inform decisions in a number of areas. For example, current methods of testing and rating water heaters, such as ASHRAE Standard 118.2 (ASHRAE 2006) and the U.S. Department of Energy's Energy Factor (DOE 2001), rate water heaters based on a 24-hour simulated usage pattern. The usage pattern consists of six equal draws of about ten gallons at three gallons per minute, spaced an hour apart. This usage pattern is not representative of actual usage and introduces some biases into water heater ratings (Schoenbauer, Hewett, and Bohac 2010). An improved hot water usage profile would make ratings more representative of actual performance and would provide for more accurate comparison, allowing homeowners to make better choices. Similarly, the actual water and energy savings from low flow fixtures is hard to determine without actual field data to determine behavioral changes due to the low flow fixtures. Knowing how often simultaneous uses actually occur and understanding the impact of pipe sizing on delay times and delivered efficiency for short draws might help shape distribution piping sizing in plumbing codes.


In 2009 and 2010, ten homes were monitored for twelve to fifteen months each to test the in situ performance of a variety of residential water heaters. Each home had two or three water heaters installed in parallel. Eight natural draft storage type water heaters (StWH), nine non-condensing tankless water heaters (NTWH), and seven condensing tankless water heaters (CTWH) were installed. The water heaters were alternated monthly in each home, and input, output, and water heating performance data were collected (Schoenbauer, Hewett, and Bohac 2010). Water heaters not in use were drained and isolated. To meet an important secondary objective, extensive data were collected on hot water draw patterns and end uses.


Each home was equipped with two water meters, a total house water meter, and a meter mounted on the cold water line into the parallel water heaters. Matched immersion resistance temperature detectors were installed at the inlet and outlet of each water heater. A surface mount thermocouple was attached to the branch of the hot water distribution system serving each hot water fixture. These thermocouples were used to indicate hot water flow to the fixture. Each thermo-couple was placed as close to the water heater as possible but on a branch that supplied only one fixture. Thermocouples were moved far enough away from tees that thermal conduction along the metal pipe would not create false readings. Figure 1 shows the typical placement of end use thermocouples. Most thermocouples were wired to a central data logger in each home that took measurements once a second and uploaded data to a central server daily. In the case of second-floor bathrooms where the plumbing was enclosed in walls, miniature data loggers with integral temperature sensors were used. The data sets from the central and remote loggers were combined during analysis. Table 1 shows the important characteristics of the data collection apparatus.

Table 1. Important Instrumentation Specifications

Instrument        Parameter     Resolution  Precision     Range

Water Meter     Total and     198.4       2% of       0.5 to 25
(positive       Water Heater  pulses/gal  readings    gpm (2 to
displacement)   Water         (53.4                   95 L/min)
                Volumes       pulses/L)               (1)

Immersion RTDs  Water Heater              1/10 DIN    -148
(matched)       Inlet and                             [degrees]
                Outlet                                F to 752
                Temperatures                          [degrees]
                                                      F (-100
                                                      C to 400

Surface Mount   Hot Flow to               Greater of  -250
Thermocouples   Fixture                   1           [degrees]
                                          [degrees]   F to 500
                                          C or        [degrees]
                                          0.75%       F (-328
                                                      C to 260

The range of the meter was selected to meet accuracy specifications.
The meter register flow repeatedly at much lower flow rates, but
the accuracy increase to around 5%.

Site Selection

A convenience sample of ten sites was selected with a mix of household sizes matching that of the single-family detached housing stock in the 2000 United States Census for Minnesota (US Census Bureau 2000). In addition, space to install at least two water heaters and sufficiently accessible piping to enable the hot water pipe temperature at each fixture to be monitored were necessary. Sites with city water were given preference over sites with private wells. The city water in the Minneapolis/St. Paul metropolitan area leaves the water treatment plants with hardness below the maximum levels recommended by tankless water heater (TWH) manufacturers. Water from private wells in the metro area is quite hard, which requires that homeowners consistently maintain their own water softening equipment to keep hardness below recommended maxi-mums. Water heaters were sized based on manufacturers' and the installing plumber's sizing guidelines and installed in conformance with manufacturers' requirements. Both plumbers and manufacturers sized heaters based on the size of the homes and the number of fixtures to prevent the possibility of an undersized water heater when the home is sold. Homes were also pre-screened for extreme water consumption behavior. However, no homes were actually excluded due to this screening. Table 2 shows the breakdown of installations.
Table 2. Water Heater Installations by Site

Site  Household Size, No.  StWH  NTWH  CTWH
           of People

1                       3  Yes   Yes   Yes
2                       4  Yes   No    Yes
3                       3  No    Yes   No
4                       2  Yes   Yes   Yes
5                       1  Yes   Yes   No
6                       5  Yes   Yes   Yes
7                       2  No    Yes   Yes
8                       4  Yes   Yes   Yes
9                       1  Yes   Yes   No
10                      2  Yes   Yes   Yes

Data were collected between December 2008 and June 2010. An alternating mode test procedure was employed. Valves were set at each site so that gas and water would flow to only one water heater at a time. Each site alternated modes, or changed water heaters, every month. This test method allowed data to be collected for each water heater under similar entering water temperature conditions and seasonal usage patterns. The schedule for water heater changeover was monitored and adjusted at each site so that every heater operated over the full spectrum of incoming water temperatures and outdoor air temperatures. At changeover, the heater being brought on line was flushed and refilled and the period required to bring it back to temperature was discarded from the data set.

Data Analysis

The one-second data were compiled into a hot water draw database. The measurement from the flow meter installed at the inlet of the water heater was used to group the one-second data into hot water events. A non-zero reading on the flow meter triggered the start of a draw, and data were summed until the next second with zero flow. Draw volume, length, flow rate, hot water energy output, and gas and electrical consumption were computed for each draw. For each draw the one-second end use thermocouple data were also scanned for increases in temperature. The rate of change of the end use temperatures was used to assign hot water draws to different fixtures. This methodology was first developed by Tiller et al. in 2004. The methodology described by Tiller et al. was applied to the data collected in this project with several modifications to meet the specific needs of this project. A 60-second backward window was used to determine the rate of change of the end use temperatures, but the criteria for determining if a rate of change was significant needed to be changed for some fixtures. The main branches and trunks of the distribution systems were copper in all ten homes; however, plastic tubing and other plumbing materials were used near the end uses at some sites. These different piping materials affect the heat conduction through the pipe to the thermocouple and required a different rate of change criteria. The criteria also needed to be adjusted in cases where two fixtures were only separated by a small distance. In these cases it was common for heat conduction along the pipe to increase the temperature reading from the thermocouple of an end use that was not active.

Draws were broken down into five categories based on their end use assignment:

* The LT5 (less than 5) category includes all draws that were less than five seconds in length. These short draws were numerous but did not contribute substantially to the total daily hot water or energy use. LT5 draws were handled separately from the rest of the analysis and were often excluded.

* The No_WH_Rise category contains all draws that were longer than 5 seconds but not long enough for hot water to exit the water heater. These draws only occurred with TWHs and were typically either very short or below the minimum firing rate for the TWH. The StWHs never took longer than 5 seconds to produce hot water.

* The No_End_Rise category includes draws where hot water came out of the water heater but hot water never made it to the fixture and consequentially no thermocouple was triggered. The time it takes for hot water to reach a fixture depends on the pipe size and length, the water flow rate, and the length of time since the last draw.

* Draws were classified as Single_End draws when only one end use thermocouple showed an increased temperature; i.e., when only one hot water use was active.

* Draws were classified as Multi_End when a hot water event triggered two or more end use thermocouples; i.e., when more than one hot water use was active.

Each of the draw types was analyzed differently because each affected hot water system performance and usage differently. Single_ End draws, which accounted for the majority of the hot water volume at all sites, were analyzed to evaluate the characteristics and usage of specific fixtures within the homes. Multi_End draws were analyzed to identify pairs of fixtures that are commonly used together and conditions that increase the likelihood of simultaneous draws. While No_WH_Rise and No_End_Rise draws do not provide hot water to a fixture, they do contribute heat to the water heater and/or the distribution system. These draws were analyzed to determine the amount of that heat that cooled in the system.


Daily Hot Water Usage and Draws

Hot water usage varied considerably from house to house and from day to day within the same house. The average daily volume of hot water consumed by all houses in this project was 37.8 gallons (143 liters), ranging from 19.6 gallons (74 liters) in the lowest use home to 59.2 gallons (224 liters) in the highest use home. The standard deviation of daily consumption at individual homes ranged from 11.6 to 34.9 gallons (44 to 132 liters). The coefficient of variance, the ratio of standard deviation to average value, ranged from 46 to 91%. These large standard deviations and coefficients of variance show the large variations in how much hot water was used from day to day. It was not unlikely that a home could use 46 to 91% more hot water from day to day. It has been theorized that people might use different amounts of hot water with a TWH than with a StWH, due to the essentially unlimited capacity, the longer time to deliver hot water, or the inability of low flows to trigger the burner. But none of these ten homes showed a statistically significant difference in total daily usage with different types of water heater. There were some differences in how hot water was used but not in the total daily volume. Figure 2 shows the cumulative frequency distributions for daily hot water consumption at each site. Figure 2 shows that some homes had similar daily hot water usage volume distributions. Three homes (Sites 5 and 9, the single-occupant homes, and Site 4, one of the two-occupant homes) used less than 20 gallons on about half of the days, while the remaining homes (including the other two-person homes) used 20 gallons or less on fewer than 10% of days. Household size was not necessarily the most important characteristic in defining the daily consumption distribution.

The number of draws per day also varied greatly from site to site and day to day considering that only draws five seconds or longer an average of 26 draws per day were taken through-out the study. Averages at individual sites ranged from 8 to 48 draws per day, with standard deviations ranging from 5 to 25. The average coefficient of variance was 69%. Table 3 shows the averages and standard deviations in daily hot water consumption and draws per day for each site. Daily hot water volume and draws generally increased with the number of occupants, but occupant behavior also played a role. For example, Site 2, with four people, used more hot water on average than Site 6 with five people.
Table 3. Daily Average Hot Water Usage and Draws for Each Site Hot
Water Gallons per Day (Liters per Day)

           Site 1  Site 2  Site 3  Site  Site  Site 6  Site 7
                                     4       5

Average      37.8    34.9    22.6  20.4  11.6    29.3    19.6
            (143)   (132)    (86)  (77)  (44)   (111)    (74)

Standard     31.0    59.2    48.7  22.3  20.4    58.8    37.3
Deviation   (117)   (224)   (184)  (84)  (77)   (223)   (141)

CV            82%     59%     46%   91%   57%     50%     53%

           Site 8  Site 9  Site    Avg

Average      32.7    16.0   18.6   21.6
            (124)    (61)   (70)   (82)

Standard     29.2    19.6   35.9   37.8
Deviation   (110)    (74)  (136)  (143)

CV            89%     82%    52%    61%

Hot Water Draws per Day (Excluding Draws Less Than 5 Seconds)

           Site  Site  Site  Site  Site  Site  Site  Site  Site  Site
              1     2     3     4     5     6     7     8     9    10

Average      20    48    35     8    16    28    37    17    20    17

Standard     18    25    26     8     5    19    22    23    18    15

CV          89%   51%   75%   90%   29%   67%   61%  135%   92%   88%


Average     26

Standard    17

CV         69%

Hot Water Usage by Draw Category

Figure 3 shows the breakdown of total draw volume by draw category for each site. For most sites, Single_End draws accounted for the largest volume of hot water by far, typically over 70%. Sites 2 and 9 had slightly lower percentages of Single_End draw volume. At Site 2, a three-occupant home, Multi_End draws had a higher percentage of total volume than at the other nine sites. Site 9 had the shortest draws of the ten homes and so had more usage in the LT5 and No_WH_Rise categories than other sites.


LT5 Draws. At all sites LT5 draws accounted for 52 to 98% of draws but only 1 to 14% of total volume (Table 4). These draws were broken down by water heater type as well. There was a significant difference in the numbers of this type of draw for different water heater types. LT5 draws were more common with StWHs than TWHS. It appeared that occupants tended to group several small draws into a single larger draw or choose not to use cold water for small draw needs when using a TWH. This change in behavior was likely due to the operational differences in TWHs and StWHs. There were an especially high number of these small draws at Site 6 due to a leak in a hot water fixture at this home. Excluding Site 6, an average of 88% of all draws were LT5 draws, but these accounted for only 5% of the hot water volume. Since these small draws do not contribute substantially to the total volume (or energy consumption), they were often excluded from the analysis. Some LT5 draws may result from water movement and pressure changes within the hot water distribution system and not actual draws. It was difficult to distinguish between these types of events and very short draws, which are common in homes.
Table 4. Characteristics of LT5 Draws at Each Site

      Number of    LT5           % of            LT5,         % of
      Occupants  Drawsper        Total           g/day        Total
                   Day           Draws         (L/day)        Volume
                   StWH    TWH   StWH   TWH    StWH    TWH    StWH

Site          3        57    36    63%  57%   1.0 (4)   0.8      2%
1                                                       (4)

Site          4       163    54    76%  56%  2.7 (10)   1.6      5%
2                                                       (6)

Site          3        NA    98     NA  74%        NA   1.2      NA
3                                                       (5)

Site          2        78    12    89%  59%   0.9 (3)   0.4      4%
4                                                       (2)

Site          1        56    30    77%  65%   1.1 (4)   0.6      5%
5                                                       (2)

Site          5      1063  1576    98%  97%  6.4 (24)   6.0     11%
6 *                                                     (23)

Site          2        NA    36     NA  52%        NA   1.1      NA
7                                                       (4)

Site          4       122    26    92%  95%     0 (0)   0.4      0%
8                                                       (1)

Site          1       330    53    94%  75%  2.7 (10)   0.8     14%
9                                                       (3)

Site          2        48    27    76%  44%   0.5 (2)   0.9      2%
10                                                      (3)


Site   2%

Site   3%

Site   2%

Site   2%

Site   3%

Site  10%
6 *

Site   3%

Site   1%

Site   4%

Site   1%

* Site 6 was found to have a leaky fixture, accounting for the very
high number of LT5 draws.

No_WH_Rise and No_End_Rise Draws. These two draw categories are grouped together because they are similar. Both add some heat to the water heater and/or the hot water distribution system but do not produce hot water at any fixture. No_WH_Rise draws occurred only with TWHs. Almost all LT5 draws had no water heater outlet temperature increase and could be classified as No_WH_Rise draws, but a special category was created for draws less than 5 seconds. These draws added heat to the heat exchanger inside the TWH but were not long enough to bring the entire unit up to temperature and produce hot water. No_End_Rise draws occurred with both types of water heater. These draws produced hot water that left the water heater and moved into the distribution pipes but did not reach the fixture. Such hot water stays in the pipes until it cools to ambient temperatures or until another draw occurs along the same branch of pipe and the hot water is used. Figure 4 shows the frequency distribution of No_WH_Rise and No_End_Rise draws by the length of time from the end of that draw to the start of a draw where hot water reached an end use fixture. Hiller (1998) found that with copper pipe it takes about 15 minutes for hot water to cool to 105 [degrees] F when the pipe is uninsulated and about an hour when insulated. Hot water energy output was summed for No_WH_Rise and No_End_Rise draws that were followed within 15 or 60 minutes by a draw that supplied hot water to a fixture. This analysis suggested that 42% of the energy output from draws of these two types could potentially be recovered on the next draw with annulated pipes. This number increases to 62% for insulated pipes. These percentages represent the maximum possible energy that could be recovered. If a draw was followed within 15 minutes by another draw that got to a fixture, before the draw got there the water in the pipes had cooled some amount and some of the heat was lost. Additionally, some of the following draws within 15 minutes will occur on a different branch of the distribution system than the previous draw. In this case only the energy in the water in shared branches will be recovered.


Single End Draws. Single_End draws accounted for 8% of all draws but 71% of draws five seconds or greater. They also accounted for 80% of the total hot water volume used. Table 6 shows the percentage of total hot water volume used that was used by Single_End draws at each fixture. Site draws do not total to 100% because draws other than Single_End are not included. Only the percentage of hot water volume that actually reached a single fixture was accounted for. Showers and tubs typically account for the most hot water use, followed by kitchen sinks. However, at two sites the kitchen sinks accounted for more hot water use than showers and baths. At Sites 5 and 10 bathroom sinks used more hot water than kitchen sinks. Laundry sinks accounted for the least hot water use in all ten homes. Table 7 shows the percentage of Single_End draws used by each fixture. Kitchen sinks were the fixture most frequently used for Single_End draws at each of the ten sites, accounting for an average of 47% of the draws for this type. Counts for clothes washers and dryers need to be considered carefully. Each water-drawing event was counted as an individual draw. The number of events per load varied considerably, and it was not practicable to group all the draws associated with a given load.
Table 5. Percentage of Hot Water Energy Output of No_WH_Rise and
No_End_Rise Draws That Could Potentially Be Used by the Next Draw

         Uninsulated Pipe  Insulated Pipe

Site 1                55%             70%
Site 2                58%             79%
Site 3                43%             64%
Site 4                26%             42%
Site 5                33%             60%
Site 6                46%             73%
Site 7                29%             49%
Site 8                48%             70%
Site 9                35%             51%
Site 10               44%             61%

Table 6. Percentage of Total Daily Volume

             Site  Site  Site  Site  Site  Site  Site
               1     2     3    4     5     6     7

Showers and   59%   23%   49%   38%   63%   43%   26%

Bath Sinks     2%    3%   14%    4%   19%    8%   20%

Kitchen        8%   29%   14%   23%    8%   17%   19%

Laundry        0%    0%    0%    3%    0%    0%    3%

Clothes        4%    6%    5%   15%    0%    3%    8%

Dish           3%    NA    NA   NA*    NA    4%    4%

Total         76%   61%   81%   83%   91%   75%   80%

             Site  Site  Site  Avg
               8       9      10

Showers and   43%   16%   70%  43%

Bath Sinks     1%   11%    8%   9%

Kitchen        7%   28%    6%  16%

Laundry        0%    0%    2%   2%

Clothes        6%    8%    0%   7%

Dish           7%    NA    NA   4%

Total         64%   63%   80%  80%

Site 4 had a portable dishwasher that attached to the kitchen sink
when in use. Therefore dishwasher draws were classified as kitchen
sink uses.

Table 7. Percentage of Single End Draws per Fixture Type

         Site  Site  Site  Site  Site  Site  Site  Site
           1     2     3     4     5     6     7     8

Showers   20%    4%    7%   14%    5%   10%    4%   11%

Bath      19%   12%   25%   16%   61%   32%   26%    7%

Kitchen   38%   73%   42%   62%   34%   48%   59%   28%

Laundry    0%    0%    0%    3%    0%    0%    3%    0%

Clothes   15%   11%   17%    5%    0%    1%    2%   28%

Dish       8%    NA    NA    NA    NA    1%    2%   26%
Washers                     (1)

         Site  Site  Avg
           9      10

Showers    4%   16%  10%

Bath      27%   28%  25%

Kitchen   67%   26%  47%

Laundry    0%    2%   2%

Clothes    3%   28%  12%

Dish       NA    NA   4%

Site 4 had a portable dishwasher that attached to the kitchen sink when in use. Therefore dishwasher draws were classified as kitchen sink uses.

Shower and tub draws account for 43% of the total volume of hot water used in the study. Shower usage was strongly dependent on occupant behavior. Overall, 0.5 showers or baths were taken per person per day, but the average for individual homes ranged from 0.3 to 1.0 showers/person/day. Showers and baths averaged 7.5 minutes long and used about 12 gallons of hot water. Table 8 shows the average shower and bath characteristics at each site. Several sites had multiple showers and/or bathtubs. The secondary fixture at Site 7 was a bathtub only, which explained the nearly 7 gpm flow rate.
Table 8. Average Shower and Bathtub Draw Characteristics

             No.     Length,  Volume,     Flow       No. of
             per       min    gal (L)   rate, gpm  Occupants
             Day                           (L/min)

Site 1 -      1.9      6.2  6.1 (23)    1.0 (4)          3

Site 1-       1.2      7.1      15.1    2.1 (8)
Primary                         (57)

Site 2 -      0.8      6.1      13.9    2.1 (8)          4
Primary                         (53)

Site 2-       0.4      8.0  6.7 (25)    0.8 (3)

Site 3        1.7      8.2      13.5    1.6 (6)          3

Site 4        0.8      8.8      10.4    1.2 (5)          2

Site 5        0.6     11.8      20.1    1.8 (7)          1

Site 6 -      0.6      4.6  5.6 (21)    1.2 (5)          5

Site 6 -      1.6      8.7      13.6    1.7 (6)
Secondary                       (52)

Site 7 -      0.3      5.7      12.1    2.1 (8)          2
Primary                         (46)

Site 7 -      0.8      5.3      12.3   6.8 (26)
Secondary                       (47)

Site 8 -      1.0     12.2      12.8    1.0 (4)
Primary                         (49)

Site 8 -      0.2      5.9  7.9 (30)    1.3 (5)          4

Site 8 -      0.3     16.5      22.6    1.4 (5)
Tertiary                        (86)

Site 9        0.5      5.5       7.4        1.6          1

Site 10       2.0      9.9      12.6    1.2 (5)          2

           No. per
           Day Per

Site 1 -       1.0

Site 1-

Site 2 -       0.3

Site 2-

Site 3         0.5

Site 4         0.6

Site 5         1.8

Site 6 -       0.4

Site 6 -

Site 7 -       0.6

Site 7 -

Site 8 -

Site 8 -       0.4

Site 8 -

Site 9         1.6

Site 10        0.6

Table 9. Average Kitchen Sink Draw Characteristics (Single_End
Draws Only)

        No. of   No. per  Length,  Volume,    Flow
      Occupants  Day per     s     gal (L)   Rate,
                 Person                      (L/min)

Site          3      2.2       30  0.6 (2)  1.1 (4)

Site          4      5.4       32  0.8 (3)  1.4 (5)

Site          3      3.6       27  0.6 (2)  1.4 (5)

Site          2      1.9       39  1.3 (5)  2.0 (8)

Site          1      4.4       16  0.4 (2)  1.5 (6)

Site          5      2.2       45  1.0 (4)  1.3 (5)

Site          2      7.7       25  0.4 (3)  1.0 (4)

Site          4

Site          1      6.7       38  0.9 (3)  1.4 (5)

Site          2      1.7       38  0.7 (3)  1.1 (4)

Table 10. Average Bathroom Sink Draw Characteristics (Single_End
Draws Only)

        No. of   No. per  Length,  Volume,    Flow
      Occupants  Day per     s     gal (L)  Rate, gpm
                 Person                          (L/min)

Site          3

Site          4      0.9       21  0.5 (2)    1.3 (5)

Site          3      2.1       43  1.1 (4)    1.4 (5)

Site          2      0.5       38  0.9 (3)    1.4 (5)

Site          1      7.6       26  0.5 (2)    1.1 (4)

Site          5      1.4       36  0.7 (3)    1.1 (4)

Site          2      3.5       38   1.1(4)    1.2 (5)

Site          4      0.2       37  0.7 (3)    1.2 (4)

Site          1      2.7       54  1.1 (4)    1.3 (5)

Site          2      1.8       49  0.8 (4)    1.0 (4)

Tables 9 and 10 list the average characteristics of kitchen sink and bathroom sink Single_End draws. The main difference between kitchen sink draws and bathroom sink draws was the number of draws per day. There were about twice as many Single_End kitchen sink draws per day (about 10), as there were bathroom sink draws (about 5). Most of the LT5, No_WH_Rise, and No_End_Rise draws are likely sink draws that were not included in Tables 9 and 10. These draws could not be included because hot water never reached the fixture, so which fixture was used was unknown.

Multi_End Draws. Multi_End draws accounted for only 2% of draws five seconds or greater and 11% of the total hot water volume. This represents a fairly small fraction of hot water behavior. As expected, the four- and five-occupant homes had more simultaneous draws, around 15% of total volume, than the single-occupant homes, where less than 5% total hot water volume was used for simultaneous uses on average. The kitchen sink was the fixture most frequently used simultaneously with other draws. Over half of the Multi_End draws involved kitchen sinks, 1.1% of all draws greater than 5 seconds. These draws account for 5% of the total hot water volume. Volume was not separated to different end uses for Multi_End volume. So the 5% contains both the water used at the kitchen sink and the water used at the other fixture simultaneously.

Hot Water Draw Characteristics

Four key characteristics of hot water draws were studied: draw length, draw volume, flow rate, and the time between the start of the draw and the end of the previous draw. For No_WH_Rise and No_End_Rise draws the idle time after a draw was significant because the analysis was looking at the potential to recover the energy from the No_WH_Rise or No_End_Rise draw. Here the idle time prior to the draw was determined because that time relates to the energy still remaining in the TWH. The idle time between draws is directly related to the cooling of the hot water remaining in the pipes after a draw. Draws that are closely spaced can take advantage of the hot water left in the system from the previous draw. Figures 5, 6, 7, and 8 show the frequency distribution of draws by length, volume, flow rate, and idle time, respectively. All four of these figures exclude draws less than five seconds.

Nearly 70% of draws were less than thirty seconds in length, and about 95% of draws were less than two minutes (Figure 5). Around 90% of draws were less than half a gallon (2 liters) in volume (Figure 6). Since the vast majority of draws are small and since total draw time is small, transient and cyclical heater performance and standby loss are important to consider as well as the steady-state in determining average efficiency. Hot water was actively being drawn for only 29 minutes per day on average. That means the water heaters were operating in standby mode for 23.5 hours per day, 98% of the time.


Very few draws were less than 0.5 gallons per minute, the typical minimum rate required for a TWH's burner to fire (Figure 7). TWHs were used at low flow rates less frequently than StWHs, suggesting that homeowners learned that they needed to use higher flow rates to get hot water, or that they would get hot water sooner with higher flow rates, when using TWHs.

Previous analysis addressed the potential energy to be recovered from No_WH_Rise and No_End_Rise draws. After all draws there was hot water left in the pipes. There is potential to recover energy left in the hot water in the pipes after all draws. Figure 8 shows that about 60% of draws occurred within 15 minutes of another draw. Insulated pipes would lengthen the cool-down time to around 60 minutes and increase the percentage of draws that could use some of the residual heat in the piping to about 80% of draws.


Effects of Hot Water Draw Types on Energy Consumption

For StWHs it is impossible to allocate energy input to individual draws because energy is continually added to or withdrawn from storage, rather than being coincident with individual hot water events. The three draw categories that comprise the smaller draw types, LT5, No_WH_Rise, and No_End_Rise, accounted for 5%, 3%, and 3% of the total hot water energy output from StWHs, compared with 7%, 2%, and 4% of the total output water volume.

The energy consumption associated with an individual draw can be determined for TWHs because the energy usage is coincident with hot water usage. The LT5, No_WH_Rise, and No_End_Rise accounted for 1%, 2%, and 3% of the total energy consumed by these heaters to heat water. There was often energy consumed to warm up the heat exchanger even with small draws without a temperature increase at the WH outlet. These draw types account for 11% of the total hot water draw volume but only 7% of the energy consumption. This was an unexpected result. It was assumed that because smaller draws were less efficient, 61% for TWH LT5, No_WH_Rise, and No_End_Rise compared to 77% for Single_End and Multi_End, they would account for a greater percentage of energy consumption than of total hot water volume. This proved not to be true because 80% of TWH LT5, No_WH_Rise, and No_End_Rise draws did not cause the burner to fire and therefore had no input. The remaining 20% of the LT5, No_WH_Rise, and No_End_Rise had gas consumption and hot water energy output, thus allowing an efficiency for these draws to be calculated, but the total energy output for these draws was small. Thus, these small draw types did not have a large impact on daily efficiency. Transient heat loss associated with heating and cooling of the TWH mass does have an important effect on daily efficiency, but mostly in the context of the Single_End and Multi_End draws. Figure 9 shows the efficiency as a function of draw length for Single_End draws at all sites. For Single_End draws, the average efficiency ranges from 67% for draws less than 30 seconds to 84% for draws greater than 2 minutes. Single_End draws account for 84% of the TWH energy consumption, so the length of these draws significantly affects daily efficiency. As a result of these transient effects, TWH daily efficiency depends on daily draw volume, as shown in Schoenbauer et al. (2010). StWH daily efficiency is even more strongly dependent on daily draw volume (Schoenbauer et al. 2010), but for these heaters this is due primarily to heat loss from the stored volume of hot water and only secondarily to transient effects.


Hourly Hot Water Usage Profile

Hot water was used at different times of the day on week-days than on weekends (Figures 10 and 11). Weekdays typically had a morning peak, often two to four times greater than any other hour of the day, an extended stretch of little or no use during the day, and then an evening peak around dinner time. Weekends had a morning peak that was later than the weekday peak. Hot water usage was lower during mid-day but higher than the same period of time on the weekdays. The evening usage peaks for weekends and weekdays were similar. The morning peak was typically associated with the shower times of the occupants. Sites with lower percentages of total volume assigned to showers and baths (Sites 9 and 2, Table 6), have much smaller morning peaks (Figure 10).


There were a high number of draws, 91%, where hot water never reached the fixture. This meant a large distribution of energy loss in homes. This energy could be saved through better plumbing design. Pipe insulation, more centrally located fixtures requiring shorter pipe runs, multiple smaller water heaters, demand-controlled recirculation loops, or small pipe diameter are several different plumbing techniques that could potentially save energy by improving distribution.

The remaining 9% of draws, for which hot water arrived at a fixture, represent 88% of the total hot water volume used and 91% of the total energy consumed. Most of this hot water went to showers, kitchen sinks, and bathroom sinks, 37%, 20%, and 11% of the total hot water volume, respectively. Hot water end use data can be used to estimate the impact of certain conservation measures. For example, if low flow showerheads could be installed so that each shower used 1 gallon per minute of hot water, which represents a reduction in all but one shower in this project, 1574 gallons of hot water could be saved per home per year. This would be a 36% reduction in shower hot water usage and a 1750 kBtu energy usage reduction assuming the home was using a StWH.

An average daily hot water consumption of 38 gallons per day was recorded for the ten houses monitored. This consumption rate meant that about 1000 Btu/hr of hot water energy were demanded by each home on a daily basis. StWHs installed in this project had an input rating of 40,000 Btu/hr, and the TWHs had input ratings ranging from 30,000 to 199,900 Btu/hr. This large difference suggests that using a water heater to provide the space and water heating needs of a home has the potential to take advantage of the maximum capacity of the appliance.


Hot water usage had a large variance even within a single home, let alone across homes. The coefficient of variation of daily hot water consumption was between 50% and 90%. The coefficient of variation of the number of draws per day was also high, between 29% and 92%. Large values for the coefficient of variance meant that there was a large degree of instability from day to day.

The number of draws per day was much higher than expected. This was due to the large number of very short draws. 88% of all draws were less than 5 seconds, and 91% of all draws never had hot water reach the fixture. The LT5, No_WH_Rise, and No_End_Rise events together accounted for about 12% of water volume and 9% of total energy consumption. Single_End draws accounted for 80% of the total hot water volume. Of Single_End draws, shower and tub draws accounted for the majority of hot water consumed, while kitchen sinks were typically the most used fixtures.

Contrary to common thinking, the ten homes in this study showed no increased daily hot water usage with tankless water heaters compared to storage water heaters. When occupants were using the TWH, they used a lower number of small draws, but their daily consumption was unchanged.
Table 11. Average Draw Characteristics per Site

           Site  Site  Site  Site  Site  Site  Site
             1    2      3     4    5     6     7

Draws per  24.1  47.8  34.6   8.5  16.2  27.7  36.8
day (no

Volume,     1.9   0.4   1.4   2.6   1.2   1.9   1.0
gal (L)     (7)   (2)   (5)  (10)   (5)   (7)   (4)

Length,    76.1  13.4  59.2  98.6  49.0  81.8  40.4

Flow        1.3   0.8   1.3   1.8   1.2   1.2   1.2
rate, gpm   (5)   (3)   (5)   (7)   (5)   (5)   (5)

Idle       0.1%  0.1%  0.0%  0.0%  0.2%  0.1%  0.1%
than 17

           Site  Site  Site    EF
             8     9    10    Test

Draws per  25.5  19.6   16.5     6
day (no

Volume,     1.8   0.9    2.1  10.7
gal (L)     (7)   (3)    (8)  (41)

Length,    97.3  38.6  103.9   214

Flow        1.1   1.2    1.2     3
rate, gpm   (4)   (5)    (5)  (11)

Idle       0.1%  0.4%   0.4%   17%
than 17

The twenty-four hour simulated usage profile used in the Energy Factor residential water heater rating assumes 63.4 gallons of hot water per day. The draw profile calls for six equal draws at one-hour intervals and a flow rate of three gallons per minute. After the final draw is completed the water heater sits in standby until the test period completes a 24-hour cycle. The Energy Factor test has a hot water usage time of 21.1 minutes.

Excluding draws less than 5 seconds, the average draw from the ten homes monitored was 1.5 gallons, 66 seconds, with a flow rate of 1.2 gpm. There were an average of 26 draws per day, and only 0.2% of draws had an idle time greater than 17 hours. This typical draw shows that the Energy Factor simulated use profile does not accurately represent how hot water is used. The simulated use pattern does not capture the cyclical and transient conditions that actual water heaters must perform under. Additionally, the measured daily hot water usage volume was 37.8 gallons per day, over 40% less than the volume assumed by the Energy Factor profile. The differences between the test procedure and actual use have the potential to greatly affect the difference between rating and actual performance.


ASHRAE. 2006. Method of Testing for Rating Residential Water Heaters.

D&R International. 2006. 2006 Buildings Energy Data Book. US Department of Energy.

DOE. 2001. Uniform Test Method for Measuring the Energy Consumption of Water Heaters.

Environmental Protection Agency. 2005. Water and Energy Savings from High Efficiency Fixtures and Appliances in Single Family Homes. US EPA-Combined Retrofit Report. March 8.

Hiller, Carl. 1998. New hot water consumption analysis and water-heating system sizing methodology. ASHRAE Transactions 104, no. 1.

Lowenstein, A., and Carl Hiller. 1998. Disaggregating residential hot water use, Part II. ASHRAE Transactions 104, no. 1.

Mayer, Peter W., and William B. Deoreo. 1999. Residential End Uses of Water. American Water Works Association, November.

Schoenbauer, Ben, Martha Hewett, and Dave Bohac. 2010. Actual Savings and Performance of Natural Gas Tank-less Water Heaters. Minneapolis, MN: Center for Energy and Environment, October 18.

Tiller, Dale, P. Gregor, D. Phil, and P. Henze. 2004. Metering Residential Hot Water By End Use: Development of Standard Protocol. 1110 South 67th Street, Peter Kiewit Institute, Omaha, Nebraska 68182: University of Nebraska-Lincoln, January.

US Census Bureau. 2000. Census 2000 Data for the State of Minnesota. US Census Bureau, April 1.

Ben Schoenbauer

Associate Member ASHRAE

Dave Bohac, PE


Martha Hewett


Ben Schoenbauer is a research engineer, Dave Bohac is Director of Indoor Air Quality, and Martha Hewett is Director of Research for the Center for Energy and Environment, Minneapolis, MN.
COPYRIGHT 2012 American Society of Heating, Refrigerating, and Air-Conditioning Engineers, Inc.
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2012 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Schoenbauer, Ben; Bohac, Dave; Hewett, Martha
Publication:ASHRAE Transactions
Article Type:Report
Geographic Code:1USA
Date:Jan 1, 2012
Previous Article:Parametric Laboratory Evaluation of Residential Heat Pump Water Heaters.
Next Article:Water and energy wasted during residential shower events: findings from a pilot field study of hot-water distribution systems.

Terms of use | Privacy policy | Copyright © 2018 Farlex, Inc. | Feedback | For webmasters