Field Performance of Heat Pump Water Heaters in the Northeast

Field Performance of

Heat Pump Water Heaters

in the Northeast

Carl Shapiro and Srikanth Puttagunta

Consortium for Advanced Residential Buildings

February 2016

NOTICE

This report was prepared as an account of work sponsored by an agency of the United States

government. Neither the United States government nor any agency thereof, nor any of their employees,

subcontractors, or affiliated partners makes any warranty, express or implied, or assumes any legal

liability or responsibility for the accuracy, completeness, or usefulness of any information, apparatus,

product, or process disclosed, or represents that its use would not infringe privately owned rights.

Reference herein to any specific commercial product, process, or service by trade name, trademark,

manufacturer, or otherwise does not necessarily constitute or imply its endorsement, recommendation, or

favoring by the United States government or any agency thereof. The views and opinions of authors

expressed herein do not necessarily state or reflect those of the United States government or any agency

thereof.

Available electronically at SciTech Connect http:/www.osti.gov/scitech

Available for a processing fee to U.S. Department of Energy

and its contractors, in paper, from:

U.S. Department of Energy

Office of Scientific and Technical Information

P.O. Box 62

Oak Ridge, TN 37831-0062

OSTI http://www.osti.gov

Phone: 865.576.8401

Fax: 865.576.5728

Email: [email protected]

Available for sale to the public, in paper, from:

U.S. Department of Commerce

National Technical Information Service

5301 Shawnee Road

Alexandria, VA 22312

NTIS http://www.ntis.gov

Phone: 800.553.6847 or 703.605.6000

Fax: 703.605.6900

Email: [email protected]

iii

Field Performance of Heat Pump Water Heaters in the Northeast

Prepared for:

The National Renewable Energy Laboratory

On behalf of the U.S. Department of Energy’s Building America Program

Office of Energy Efficiency and Renewable Energy

15013 Denver West Parkway

Golden, CO 80401

NREL Contract No. DE-AC36-08GO28308

Prepared by:

Carl Shapiro and Srikanth Puttagunta

Steven Winter Associates, Inc.

of the

Consortium for Advanced Residential Buildings

61 Washington Street

Norwalk, CT 06854

NREL Technical Monitor: Cheryn Metzger

Prepared under Subcontract No. KNDJ-0-40342-03

February 2016

The work presented in this report does not represent

performance of any product relative to regulated

minimum efficiency requirements.

The laboratory and/or field sites used for this work are

not certified rating test facilities. The conditions and

methods under which products were characterized for

this work differ from standard rating conditions, as

described.

Because the methods and conditions differ, the reported

results are not comparable to rated product performance

and should only be used to estimate performance under

the measured conditions.

Contents

List of Figures ............................................................................................................................................ vi

List of Tables ............................................................................................................................................. vii

Definitions ................................................................................................................................................. viii

Executive Summary ................................................................................................................................... ix

1 Introduction ........................................................................................................................................... 1

1.1 Background ...................................................................................................................................... 2

1.2 How Heat Pump Water Heaters Work ............................................................................................. 4

1.3 Space-Conditioning Interactions and Installation Considerations ................................................... 5

1.4 Model Operation and Control Logic ................................................................................................ 5

1.4.1 General Electric Control Logic ........................................................................................... 6

1.4.2 A.O. Smith Control Logic ................................................................................................... 6

1.4.3 Stiebel Eltron Control Logic ............................................................................................... 7

2 Recent Studies ...................................................................................................................................... 8

3 Technical Approach ............................................................................................................................. 9

3.1 Research Questions .......................................................................................................................... 9

3.2 Measurements .................................................................................................................................. 9

3.3 Equipment ...................................................................................................................................... 10

3.4 Analysis ......................................................................................................................................... 11

4 Performance Results .......................................................................................................................... 13

4.1 Impact of Water Use on Efficiency................................................................................................ 14

4.2 Impact of Air Temperature on Efficiency ...................................................................................... 16

4.3 Delivered Water Temperature ........................................................................................................ 19

4.4 Calculating Standby Losses ........................................................................................................... 23

4.5 Impact of Confined Spaces ............................................................................................................ 26

4.6 Impact on Space-Conditioning Systems ........................................................................................ 27

5 Energy and Cost Analysis ................................................................................................................. 30

6 Homeowner Surveys .......................................................................................................................... 35

7 Recommendations ............................................................................................................................. 36

8 Conclusion .......................................................................................................................................... 37

References ................................................................................................................................................. 40

Appendix A: A History of Heat Pump Water Heaters in the United States ......................................... 44

Appendix B: Alternative Water-Heating Calculations ........................................................................... 48

Appendix C: Monitored and Modeled Energy Use and Costs by Site ................................................. 49

Appendix D: Identifying Heating Events ................................................................................................. 51

Appendix E: Trifold Brochure for Consumers ....................................................................................... 53

List of Figures

Figure 1. Timeline of HPWH development in the United States ............................................................. 3

Figure 2. How HPWHs work ....................................................................................................................... 4

Figure 3. Example of an HPWH monitoring system installation .......................................................... 11

Figure 4. Ideal operation of the GE HPWH in hybrid mode (Site 9) ..................................................... 15

Figure 5. Low water use can reduce the overall benefit of an HPWH when the cost benefit is

assessed. ............................................................................................................................................. 15

Figure 6. Scatter plots of daily hot water use versus COP color coded by electric resistance

fraction ................................................................................................................................................. 16

Figure 7. Because the Site 5 basement was too cold, this GE HPWH switched primarily to electric

resistance mode during the winter months. .................................................................................... 17

Figure 8. Ambient temperature at Site 11 ............................................................................................... 18

Figure 9. This HPWH benefited from the waste heat of the boiler used for space heating. ............. 18

Figure 10. Typical winter performance of the GE HPWH in the Site 11 warm mechanical room ..... 18

Figure 11. Scatter plots of COPs versus daily hot water use color coded by ambient temperature.

Periods with electric resistance fractions greater than 0.04 and Sites 5 and 13 were

excluded. ............................................................................................................................................. 19

Figure 12. This Stiebel Eltron HPWH was able to maintain high COPs during high water demand. 20

Figure 13. Even with high water demand, the hot water temperature dropped only as low as

118°F. ................................................................................................................................................... 21

Figure 14. Normalized histograms of delivered water temperature. Red line indicates set point

temperature. Green line indicates minimum delivery temperature (110°F). The percentage of

occurrences with water temperatures lower than 110°F is shown. .............................................. 22

Figure 15. Regression results and data points for GE standby losses. Temperature difference is

between the tank average temperature and ambient air. ............................................................... 26

Figure 16. Performance of the HPWH at Site 13 in a confined space .................................................. 27

Figure 17. Box plots of maximum potential space-conditioning impacts for monitored HPWHs .... 28

Figure 18. Annual operating cost of monitored HPWHs and alternative water heaters .................... 31

Figure 19. Annual source energy of monitored HPWHs and alternative water heaters .................... 32

Figure 20. Annualized energy-related costs of monitored HPWHs and alternative water heaters .. 34

Figure 21. First patented HPWH. Image from Wilkes and Reed 1937. ................................................. 44

Figure 22. Results from using daily data for analysis ........................................................................... 51

Figure 23. Example of five heating events as defined in this study .................................................... 52

Unless otherwise noted, all figures were created by CARB.

vii

List of Tables

Table 1. Performance Summary of Monitored HPWHs by Model .......................................................... ix

Table 2. Residential Water-Heating Energy Consumption and Operating Costs (per annum) ........... 1

Table 3. Key Specifications of Integrated HPWHs Currently Available in the U.S. Market ................. 2

Table 4. Installed Monitoring Equipment ................................................................................................ 11

Table 5. Summary Statics of Performance by Model ............................................................................ 13

Table 6. Summary Table of Performance by Site .................................................................................. 14

Table 7. Tank Heat Loss Robust Regression Results with Constant (number of observances = 113,

degrees of freedom = 111) ................................................................................................................. 25

Table 8. Tank Heat Loss Robust Regression Results without Constant (number of observances =

113, degrees of freedom = 112) ......................................................................................................... 25

Table 9. Alternative Water Heaters .......................................................................................................... 30

Table 10. Fuel Prices and Characteristics for Energy and Cost Analysis .......................................... 31

Table 11. Cost Analysis Assumptions .................................................................................................... 32

Table 12. Installation Costs and Lifetimes of Water-Heating Products ............................................... 33

Table 13. Installed HPWH Cost Estimates from Study .......................................................................... 33

Table 14. Survey Results of Whether the Homeowner or a Qualified Professional Performed any

Preventive Maintenance Procedures ................................................................................................ 35

Table 15. Survey Results of Homeowners .............................................................................................. 35

Table 16. Summary of Test Results by Model ........................................................................................ 37

Table 17. Key Specifications of Integrated HPWHs Currently Available in the U.S. Market ............. 47

Table 18. Annual Site Energy Consumption by Site ............................................................................. 49

Table 19. Annual Source Energy (MMBtu) Consumption by Site ........................................................ 49

Table 20. Annual Operating Costs ($) by Site ........................................................................................ 50

Table 21. Annualized Energy-Related Costs ($) by Site ........................................................................ 50

Unless otherwise noted, all tables were created by CARB.

viii

Definitions

Btu

British Thermal Unit

Carbon Dioxide

COP

Coefficient of Performance

energy factor

EPRI

Electric Power Research Institute

ERWH

Electric resistance water heater

FHR

First Hour Rating

Foot

gal

Gallon

General Electric

GPD

Gallons per Day

in.

inch

Kilowatt

kWh

Kilowatt-Hour

lbm

Pound Mass

MBtu

Million British Thermal Units

HPWH

Heat Pump Water Heater

quad

Quadrillion British Thermal Units

SWA

Steven Winter Associates, Inc.

therm

100,000 British Thermal Units

Watt-Hour

Executive Summary

Heat pump water heaters (HPWHs) are finally entering the mainstream residential water heater

market. Possible catalysts are increased consumer demand for more energy-efficient electric

water heating and a new federal water-heating standard that effectively mandates the use of

HPWHs for electric storage water heaters with nominal capacities higher than 55 gal. Compared

to electric resistance water heaters (ERWHs), the energy and cost savings potential of HPWHs is

tremendous. Converting all ERWHs to HPWHs could save American consumers $7.8 billion

annually ($182 per household) in water heater operating costs and cut annual residential source

energy consumption for water heating by 0.70 quads.

Steven Winter Associates, Inc., a partner of the U.S. Department of Energy’s Building America

research team Consortium for Advanced Residential Buildings, embarked on one of the first in

situ studies of these newly released HPWH products through a partnership with two sponsoring

electric utility companies, National Grid and NSTAR, and one sponsoring energy-efficiency

service program administrator, Cape Light Compact. Recent laboratory studies have measured

the performance of HPWHs under various operating conditions, but publically available field

studies have been less available. This evaluation attempts to provide publically available field

data about new HPWHs by monitoring the performance of three recently released products:

General Electric (GE) GeoSpring, A.O. Smith Voltex, and Stiebel Eltron Accelera 300. Fourteen

HPWHs were installed in Massachusetts and Rhode Island and monitored for more than 1 year.

Of these, 10 were GE models (50-gal units), 2 were Stiebel Eltron models (80-gal units), and 2

were A.O. Smith models (1 60-gal and 1 80-gal unit).

Although this study used a small sample size and all the water heaters were in unconditioned

basements, the HPWHs studied show great promise. Excluding one site, the monitored units had

coefficients of performance (COPs) ranging from 1.5 to 2.6. The excluded site had ambient air

temperatures lower than 50°F for much of the year that resulted in excessive electric resistance

backup use. The average COP for each model is provided in Table 1.

Table 1. Performance Summary of Monitored HPWHs by Model

Model

Capacity

(gal)

First Hour

Rating

(gal/h)

Measured

Average COP

COP Range

1.82

1.5–2.1

A.O. Smith

60/80

68/84

2.12

2.1

Stiebel Eltron

78.6

2.32

2–2.6

The monitored data show that the primary variables that affect HPWH performance are hot water

use (daily volume and draw pattern) and ambient temperature. High hot water demand reduces

efficiency by increasing auxiliary electric resistance use. Higher ambient temperature increases

efficiency by increasing the efficiency of the heat pump and reducing standby losses. The GE

unit shows two distinct operating regions that correspond to large and small electric resistance

loads: the A.O. Smith and Stiebel Eltron units, which operate entirely in the low electric

resistance region. This is most likely a product of the larger tank volumes and the control logic

that allows the heat pump to reengage, in the case of A.O. Smith, or operate simultaneously, in

the case of Stiebel Eltron.

Despite the slower recovery rate of the heat pump compared to electric resistance elements, all

three models delivered hot water at temperatures higher than the minimum acceptable level

(110°F) during nearly all draws. The hybrid nature of these systems allows them to deliver hot

water reliably.

Unfortunately, the standby losses of the systems are higher than traditional ERWHs. Possible

causes are the additional piping, wraparound condensing unit, and inadequate insulation.

Installation of HPWHs in confined spaces also reduced efficiency by approximately 16%, which

is consistent with other studies.

The HPWH monitoring results were compared to several alternative natural gas, electric

resistance, fuel oil, and propane storage tank water heaters. Tankless water heaters were not

considered. With the exception of condensing storage natural gas water heaters, annual operating

costs and source energy consumption for the monitored HPWHs were lower than the alternative

storage tank water heaters considered. The annualized energy-related costs—which are a

measure of total lifetime costs and include first costs, operating costs, replacement costs, and the

time value of money—of the monitored HPWHs were slightly lower than ERWHs and

condensing natural gas water heaters. Annualized energy-related costs for HPWHs were

considerably lower than for propane- and fuel-oil-fired systems. Natural gas storage water

heaters, with the exception of condensing storage water heaters, had lower annualized energy-

related costs than all other options. Space-conditioning interactions for HPWHs, however, may

change the relative costs depending on the climate and location of the HPWH. Natural gas water

heaters, however, were still the lowest-cost storage water heater option on an annualized energy-

related cost basis.

1 Introduction

A confluence of regulatory and economic factors is rapidly pushing heat pump water heaters

(HPWHs) into the mainstream residential marketplace. The primary regulatory catalyst is a new

federal water-heating standard that mandates energy factors (EFs) around 2 for all new electric

storage water heaters with capacities higher than 55 gal (DOE 2010).

This regulation is a major

driver of change in residential water-heating technologies (Maynard 2011), because it effectively

requires HPWHs in applications that have large hot water loads and where electricity is used for

water heating. Also, for energy-conscious consumers who want to decrease energy use, HPWHs are

currently the only ENERGY STAR

-qualified electric water-heating products on the market

(EPA 2012).

In addition to a changing regulatory environment in the residential electric resistance water heater

(ERWH) market, financial factors are also pushing HPWHs into the mainstream. Inflation of

residential retail electricity prices significantly outpaced general inflation, as measured by the

consumer price index, between 2002 and 2009.

While electricity prices have since stabilized due to

a slowdown in economic growth and declining natural gas prices (EIA 2012a), the relative increase

in electricity prices over the past decade has played an important role in HPWH development.

Furthermore, even though HPWHs have higher first costs than traditional ERWHs, many utility

companies are offering sizable rebates for HPWH installations

in the hopes of making these units

more attractive in the residential marketplace.

The move to HPWHs from standard ERWHs is not trivial; energy used by electric water heaters is a

substantial fraction of total residential energy consumption, and HPWHs are significantly more

efficient than traditional ERWHs. Water heating is the second-largest contributor (2.11 quads or

20% of site energy) to residential energy consumption in the United States after space heating (EIA

2005), and nearly 44% of American households use electricity as their primary water-heating fuel

(EIA 2009). To demonstrate the magnitude of the energy consumed through residential water

heating, site energy, source energy, and annual operating costs are listed in Table 2.

Table 2. Residential Water-Heating Energy Consumption and Operating Costs (per annum)

Fuel

Number of

Households

(millions)

Site Energy

Source Energy

Operating Costs

Per

Household

U.S.

Total

(quads)

Per

Household

(MBtu)

U.S.

Total

(quads)

Per

House-

hold

($)

U.S.

Total

(billion $)

Electricity

43.3

2,813 kWh

0.42

32.3

1.40

330

14.28

Natural Gas

59.8

241 therms

1.44

26.3

1.57

265

15.80

Fuel Oil

4.3

226 gal

0.14

36.6

0.16

887

3.81

Propane

5.8

277 gal

0.15

29.2

0.17

790

4.58

Source: Adapted from EIA (2005).

The regulation specifies the minimum EF as a function of rated storage volume, EF = 2.057-0.00113V, which

corresponds to an EF range of 1.92 to 1.99 for rated tank volumes between 55 gal and 119 gal, respectively.

The average residential retail price of electricity increased 36% between 2002 and 2009 (EIA 2012b), while the

increase in the consumer price index was 19% over the same period (BLS 2012).

For example, Massachusetts utilities offer up to $1,000 (Mass Save 2012).

Because all new HPWHs have a listed EF of 2 or higher—compared to 0.9–0.95 for ERWHs

—if

all ERWHs were replaced with HPWHs and these water heaters performed at their rated

efficiencies, American consumers could save $7.8 billion annually (average of $182/household) in

water heater operating costs and cut annual residential source energy consumption for water heating

by 0.70 quads. HPWHs are not appropriate in all circumstances, however, and they may increase

space-conditioning loads in some cases, so these figures represent the upper limit of potential

savings based on EF ratings associated with the move from ERWHs to HPWHs.

In 2013, five major integrated

HPWH products were available in the American market. Key

specifications for these HPWHs are shown in Table 3. These specifications include the two U.S.

Department of Energy performance metrics—EF and first hour rating (FHR).

The EF represents

the efficiency of the electric heating elements and tank losses under a specific 24-hour test

procedure. The FHR represents the amount of hot water that can be supplied by a fully heated

storage water heater during the first hour of operation. Storage tank volume is often used as a proxy

for water-heating capacity, but FHR is the more appropriate metric. See Section 4.2 for more details

about the EF and FHR test procedures.

Table 3. Key Specifications of Integrated HPWHs Currently Available in the U.S. Market

Model

Capacity

(gal)

FHR

(gal)

Electric Resistance

Elements

(kW)

General Electric (GE)

GeoSpring

50 2.35 63

Upper: 4.5

Lower: 4.5

A.O. Smith Voltex

60/80 2.33 68/84

Upper: 4.5

Lower: 2

Stiebel Eltron Accelera 300

2.51

78.6

Upper: 1

Rheem EcoSense

40/50 2.00 56/67

Upper: 2

Lower: 2

AirGenerate AirTap

Integrated

50/66 2.39/2.40 60/75 Upper: 4

1.1 Background

Although regulatory and economic factors may be finally pushing HPWHs into the mainstream

residential market, HPWHs are not new. The first patented HPWH dates back to 1935, and they

were first commercialized in the 1950s. For more than 50 years, HPWHs have followed a boom-

bust cycle of product development and subsequent abandonment that closely follows retail

EF is meant to represent the efficiency of residential water heaters as a fraction. For example, 0.92 EF water heaters

are expected to be 92% efficient when tested under the rating procedure. See Section 4.2 for more information about the

EF test.

Integrated refers to water heaters in which the heat pump and tank are packaged together as a one-piece, drop-in unit.

Add-on HPWHs can be plumbed into existing storage water heaters.

Recovery efficiency (RE) is also used to quantify the efficiency of natural gas, fuel oil, and propane water heaters. RE

is an approximate metric of the burn efficiency of the heating device. Because electric resistance elements have an

efficiency of 1, RE is not tested or reported.

Any omission of a manufacturer or product is unintentional, and no endorsement of any commercial product or

manufacturer is implied.

electricity prices and demonstrates the reliability issues of past products. See Figure 1 for a timeline

of the development of HPWHs in the United States and Appendix A for more detailed information.

Figure 1. Timeline of HPWH development in the United States

Because of regulatory and financial forces that have converged to push HPWHs into the mainstream

residential market and the magnitude of the savings potential represented by this market, Steven

Winter Associates, Inc. (SWA), a partner of the U.S. Department of Energy’s Building America

research team Consortium for Advanced Residential Buildings, embarked on one of the first in situ

studies of these newly released HPWH products. Recent laboratory studies have quantified the

performance of HPWHs under various operating conditions, but publically available field studies

have been less available. This evaluation attempts to provide publically available field data about

new HPWHs by quantifying the field performance of three products: GE GeoSpring, A.O. Smith

Voltex, and Stiebel Eltron Accelera 300.

Fourteen HPWHs were installed in Massachusetts and Rhode Island and monitored for more than 1

year. The sites were chosen in the residential markets of the sponsoring electric utility companies,

National Grid and NSTAR, and the sponsoring energy-efficiency service program administrator,

Cape Light Compact. Of the 14 units, 10 were GE models (50-gal units), 2 were Stiebel Eltron

models (80-gal units), and 2 were A.O. Smith models (1 60-gal and 1 80-gal unit). The 10 50-gal

units were intended to be split between the GE and Rheem HPWHs, but the Rheem units were not

available during the installation phase of this project.

1.2 How Heat Pump Water Heaters Work

HPWHs are primarily designed as replacements for traditional ERWHs and can achieve higher

efficiencies by using the vapor-compression heat pump cycle. Generally speaking, heat pumps are

devices—such as air conditioners and refrigerators—that move thermal energy from one location to

another. A refrigerator moves heat from the inside the appliance into the kitchen; the heat pump

inside an HPWH moves heat from the surrounding air into the hot water storage tank. By moving

thermal energy instead of converting electricity to heat, heat pumps are more efficient and usually

operate at efficiencies that exceed 200%.

Most HPWHs, however, are hybrid devices that combine a heat pump, backup electric resistance

element(s), and a storage tank (Figure 2). Although the heat pumps in these hybrid water heaters

can heat water at high efficiencies, their recovery rate is significantly slower than traditional electric

resistance heating mechanisms. A typical 4.5-kW electric resistance element can reliably heat more

than 20 gal of water per hour. The heat pump has a lower heating rate; GE, for example, publishes a

rate of 8 gal/hour at 68°F air temperature (GE 2010). Auxiliary electric resistance elements are thus

also installed in HPWHs for reliability and quicker hot water recovery. Most HPWHs use the heat

pump whenever possible, but built-in controls switch to conventional resistance heating during

times of high hot water demand.

Figure 2. How HPWHs work

Historically, heat pumps have been differentiated into two categories: integrated (or “drop-in”) and

remote (or “add-on”). These categories describe the relationship between the heat pump and the

storage tank. Add-on devices are stand-alone heat pumps that are combined with traditional tank

ERWHs. Integrated devices include the heat pump, storage tank, and resistance element(s) in one

package and have the advantage of being able to control the operation of the heat pump and

resistance elements more precisely. Add-on devices can, however, be used as retrofits to existing

water heaters at a lower cost (Ashdown et al. 2004).

Integrated devices come in three configurations of heat exchangers. The most common in the

United States is a wraparound condenser. In this configuration, the refrigerant coils are wrapped

around the storage tank but are not in direct contact with the potable water inside the tank. The Air

Generate AirTap uses refrigerant immersion coils that are directly immersed in the potable water.

Finally, coaxial heat exchangers can be used, as with the Rheem EcoSense, wherein the water is

pumped through the heat exchanger to transfer energy from the heat pump to the potable water

(Sparn et al. 2011).

The most common arrangement of current HPWHs is an integrated water heater with a wraparound

condenser and two backup electric elements. Figure 2 illustrates this arrangement and describes the

typical components and operation of modern HPWHs in the U.S. market. Among the systems

evaluated in this study, the only model to deviate from this configuration is the Stiebel Eltron unit,

which has only one smaller (1.7-kW) upper element and always operates in hybrid mode.

1.3 Space-Conditioning Interactions and Installation Considerations

HPWHs move thermal energy from the surrounding air into the storage tank. Therefore, units

installed in conditioned spaces directly affect the space-conditioning loads of the building. During

the heating season, HPWHs increase loads on space-heating systems. Conversely, HPWHs reduce

cooling loads during the cooling season. When HPWHs are installed in unconditioned spaces such

as attics and garages (possible in warm climates only), space-conditioning impacts are minimal.

When in “quasi-conditioned” spaces—such as uninsulated basements—space-conditioning impacts

of HPWHs are very difficult to assess. Because most basements are thermally connected to

conditioned space, HPWHs still have an impact on conditioning loads. Finally, by using a vapor-

compression system, HPWHs typically remove moisture from the surrounding air, which can be a

significant benefit in hot-humid climates and damp basements.

Because these hybrid HPWHs are new to the mainstream market, installation is less straightforward

for installers than traditional ERWHs. HPWHs require special attention to airflow around the unit

and condensate collection. Furthermore, to improve the efficiency of the units, A.O. Smith and

Stiebel Eltron manufacture models that are substantially larger than typical ERWHs. Installers may

not be familiar with these units—or with heat pump models in general—and installing these units in

existing homes may be difficult.

During installation the team noted specific challenges the installer faced at the various sites. Many

of these details are discussed in more depth by Shapiro, Puttagunta, and Owens (2012). Installation

considerations are further discussed by SWA (2012). A trifold brochure for consumers (Appendix

E) was created for SWA’s rebate programs.

1.4 Model Operation and Control Logic

Although all three monitored HPWHs have similar components, the control logic differs

substantially among models. The GE and A.O. Smith tanks have two electric resistance (upper and

lower) elements in addition to the heat pump and operate in several modes; the Stiebel Eltron unit

has only one mode of operation and uses a small upper element to supplement the heat pump. All

three units have wraparound condensers and an integrated storage tank.

1.4.1 General Electric Control Logic

The GE model has two 4.5-kW electric resistance elements, one placed at the top third and one

placed at the bottom third of the unit. The unit can operate under five operating modes: Hybrid,

eHeat, Standard Electric, High Demand Mode, and Vacation Mode (GE 2009). The control logic for

these modes is described below:

• Hybrid Mode is the default operating mode. The control logic for the GE unit is

proprietary; however, the general operating principle behind the hybrid mode is to balance

energy savings and provide hot water at rates that are similar to comparably sized ERWHs.

Thus, the control logic tries to meet certain targets for the availability of hot water. The unit

uses the heat pump until approximately 75% of the available hot water has been depleted

and the heat pump cannot keep up with the hot water demand. Because the unit does not

measure average tank temperature or hot water flow rate directly and measures the

temperature at the top third of the tank only, the control logic uses the current and previous

values of the temperature at the top of the tank as a proxy for available hot water and hot

water flow rate. The system uses a closed-form model that employs autocorrelation

functions to map upper tank temperature to hot water availability (Tsai 2012).

In practice, this control strategy results in the electric elements being used in three

circumstances.

o If the ambient air temperature is outside the safe operating range (45°–120°F), the

system reverts to standard resistance mode.

o If the water in the tank is significantly lower than the set point, the upper element

operates. The difference between the tank temperature and the set point depends on

the circumstances, but it is generally 25°–30°F.

o If the system senses that the water use is too high, the lower element operates. In

general, 25–30 gal within a short time period is considered high water use. Once the

lower electric resistance element engages, the entire tank is reheated like a traditional

ERWH (Tsai 2011).

• eHeat Mode uses only the heat pump, unless the ambient temperature is outside the safe

operating range. This mode is more efficient but may fail to provide water at the set point

temperature. The temperature dead band at the top of the tank that regulates heat pump

operation is 1°F (Sparn et al. 2011).

• Standard Electric Mode operates like a traditional ERWH.

• High Demand Mode is similar to the hybrid mode, but the control logic changes from 75%

of the available hot water being depleted to 50% (Tsai 2012).

• Vacation Mode is similar to eHeat Mode but with a temperature set point of 50°F.

1.4.2 A.O. Smith Control Logic

The A.O. Smith model has two electric resistance elements: a 4.5-kW upper element and a 2-kW

lower element. The A.O. Smith model has four operating modes: hybrid mode, efficiency mode,

electric mode, and vacation mode (A.O. Smith 2010). The control logic of the A.O. Smith unit,

however, is quite different than that of the GE unit:

• Hybrid Mode of the A.O. Smith model uses a simple temperature dead-band algorithm. If

the average tank temperature—which is the weighted average of the upper and lower

thermostats, where the upper thermostat receives 75% weighting (Sparn et al. 2011)—drops

9°F below the set point, the heat pump is turned on to heat the water back to the set point. If,

however, the heat pump fails to heat the water sufficiently, and the average tank temperature

drops more than 20°F below the set point, the upper element replaces the heat pump as the

heating source. The lower element is not used in hybrid mode (A.O. Smith 2011). Unlike the

GE unit, the heat pump may reengage during the reheat cycle after the electric resistance

elements have been operating.

• Efficiency Mode does not use the electric resistance elements, unless the ambient

temperature is outside the safe operating range (45°–109°F) of the heat pump (A.O. Smith

2011).

• Electric Mode operates like a traditional ERWH. The upper element is used first to heat the

top of the tank, then the lower element is used to heat the bottom of the tank (A.O. Smith

2010).

• Vacation Mode is identical to efficiency mode with a set point of 60°F (A.O. Smith 2010).

1.4.3 Stiebel Eltron Control Logic

The Stiebel Eltron has only one operating mode. The hot water temperature set point is factory set

at 140°F and is not easily adjustable by the user. The unit has one 1.69-kW electric resistance

element installed vertically at the top of the tank and operates under a fixed mode. The heat pump is

turned on when the temperature 16 in. from the top of the internal tank drops more than 4°F below

the set point. If the heat pump cannot meet the demand and the temperature at the top of the tank

drops lower than 112°F, the upper element is used as a backup heat source. The upper element heats

only the top third of the water heater tank (approximately 27 gal). The Stiebel Eltron unit is the only

unit that allows simultaneous operation of the heat pump and booster resistance heater (Megliola

2011).

2 Recent Studies

Since the introduction of the most recent line of HPWHs, a growing body of literature has been

devoted to measuring the performance, energy and cost savings potential, and reliability of these

products.

Unfortunately, robust and publically available field-test results for the new HPWHs have been less

available than laboratory testing results. Pacific Gas & Electric monitored one GE HPWH in

Sonora, California. The unit was placed in an unconditioned basement with a floor area of

approximately 400 ft

. The average COP was 1.29 with average operating conditions of 68.6°F

ambient temperature, 62% relative humidity, 70.5°F inlet water temperature, 1.3 gal/minute flow

rate, 18.6 draw events per day, and 18.3 GPD (Hu and Davis 2011).

Electric Power Research Institute (EPRI) has worked with four electric utility companies—

Bonneville Power Administration, Kansas City Power & Light, Tennessee Valley Authority, and

Southern Company—to test 160 water heaters across the United States, mostly GE, A.O. Smith, and

Rheem HPWHs, but several ERWHs were installed as control units. EPRI has provided preliminary

results from the study, but the results are identified only as Model A, Model B, and Model C. The

COPs of these units were 0.7–2.7; their performance varied significantly. Their reliability has been

high so far, but EPRI saw only minimal peak demand savings (Amarnath and Bush 2012).

A 2009 Pacific Gas & Electric HPWH study (PG&E 2009) concluded that, in terms of source

energy

efficiency, an HPWH was more efficient (67%) than a standard natural gas (57%) or

electric (29%) tank water heater when the heating, ventilating, and air-conditioning interaction of

the HPWH is ignored. If this interaction is included, HPWHs are significantly more efficient (104%

source energy efficiency) than both standard systems in the cooling season and less efficient (44%

source energy efficiency) than natural gas tank water heaters in the heating season.

According to Franco et al. (2010), HPWHs have the lowest life cycle cost in roughly half of all

single-family homes that heat water with electricity. This study assumed that some houses need

venting for successful HPWH installation; thus, HPWHs could not be cost-effectively installed in

many older homes. Furthermore, HPWHs demonstrated a greater cost benefit in new, single-family

homes.

Hudon et al. (2012) modeled HPWH performance against gas water heaters and ERWHs in six U.S.

cities. HPWHs saved source energy compared to traditional ERWHs regardless of climate and

location of the unit (i.e. whether located in conditioned or unconditioned space). Savings over

natural gas depended on climate and location of the unit.

See Section 5 for discussion of source energy.

3 Technical Approach

This evaluation of HPWHs provides valuable information that can be used to advise consumers and

builders about efficient methods to provide electric water heating. This information can also be used

as validation by utilities and other program implementers throughout the United States that are

attempting to develop incentive programs for this technology. Although this study is primarily

applicable to colder climates, the measured performance of the HPWH units is relevant in many

climate zones. Even though the sample for this evaluation was relatively small, some clear results

were consistent with other ongoing field-testing across the country.

3.1 Research Questions

This research effort focused on answering the following questions about the efficiency, reliability,

and performance of each model evaluated:

• What is the measured efficiency of an HPWH located in unfinished basements of cold-

climate homes?

• What are the critical criteria that affect the installed performance of HPWHs, and how do

they impact the performance of each HPWH model?

• What are the standby losses for these HPWHs, and how do they compare to traditional

ERWHs?

• Does each model evaluated in this study effectively deliver hot water at the set point

temperature?

• Are homeowners satisfied with hot water delivery, efficiency, noise, and other

characteristics?

3.2 Measurements

Long-term performance data were collected at 14 sites in Massachusetts and Rhode Island in the

service districts of National Grid, NSTAR, and Cape Light Compact. Measurements were taken for

a minimum of 12 months at all sites to establish the annual efficiency and performance of each unit.

These measurements included water temperatures, flow rates, and electricity consumption. Sensors

were sampled at 5-second intervals and output at 15-minute intervals in the forms of averages,

minimums, maximums, and/or totals over that time period, depending on the desired outputs. An

additional data table captures the duration and volume of each hot water draw.

At each site, the following HPWH parameters were measured every 5 seconds:

• Inlet water temperature (°F)

• Outlet water temperature (°F)

• Ambient air temperature (°F)

• Ambient air relative humidity (%)

• Hot water flow (gal)

• Compressor energy consumption (Wh)

• Energy consumption of each electric resistance heating element (Wh)

• Entire system energy consumption (Wh).

At each 5-second interval, thermal heat delivered by the water heater was calculated based on water

flow and temperature differential.

The following values were output for each 15-minute logging period:

• Average water inlet temperature (°F)

• Average water outlet temperature (°F)

• Minimum water inlet temperature (°F)

• Maximum water outlet temperature (°F)

• Average inlet air temperature (°F)

• Average inlet air relative humidity (%)

• Total domestic hot water use (gal)

• Domestic hot water energy (Btu)

• Total compressor energy consumption (Wh)

• Total upper heating element energy consumption (Wh)

• Total lower heating element energy consumption (Wh)

• Total system energy consumption (Wh)

• Total heat pump energy (Wh)

• Total standby energy consumption (Wh)

The start time, volume, and duration of each hot water draw were also recorded. Details on how a

heating event was defined for this study is provided in Appendix D

3.3 Equipment

As shown in Figure 3, a Campbell-Scientific CR-1000 data logger and various sensors (Table 4)

were used at each site to take measurements. A wireless modem was used to remotely download

data for evaluation throughout the monitoring period. Thermistors were used for all temperature

measurements. Cold water inlet and hot water outlet water temperatures were measured using an

Omega tubular immersion sensor with a 4.5-in. probe length and National Pipe Thread tapers. The

HPWH installer installed these directly in the water flow. Omega FTB4607 low-flow (0.22–20

gal/minute) turbine-type flow meters were used to measure domestic hot water flow. All flow

meters were installed by the HPWH installer and located on the cold water inlet side.

Air temperature and relative humidity in the space surrounding the HPWH were measured by a

Humirel HTM2500 located to minimize heat transfer from radiation and surrounding equipment.

All electrical energy consumption measurements used a Continental Control Systems WattNode and

right-sized current transformers. The WattNodes are true-root mean squared, alternating current,

watt‐hour transducers with pulse outputs.

Table 4. Installed Monitoring Equipment

Measurement

Equipment

Record and Output Measurements

Campbell-Scientific CR-1000 data logger

Inlet and Outlet Water Temperatures

Omega ON-910-44006 National Pipe Thread pipe

plug thermistor

Inlet Air Temperature and Relative

Humidity

Humirel HTM2500 Probe

Compressor Energy, Heating Element

Energy, and System Energy

Continental Control Systems WattNode WNB-3D-

240-P watt-hour transducer with appropriately sized

current transformers

Hot Water Flow

Omega FTB4607 low-flow, turbine-type flow meter

Figure 3. Example of an HPWH monitoring system installation

3.4 Analysis

The COP has been defined differently in numerous studies (AIL Research 2001, AIL Research

2002, Murphy and Tomlinson 2002, and Zogg 2002). For this evaluation the standard definition of

COP, which is the net heat delivered by the water heater to the domestic water load divided by the

total electrical energy consumed over a period of time was used:

Btu/Wh 413.3inputenergy net

energy heating useful

DHW

COP

(1)

where

COP

= coefficient of performance (dimensionless)

DHW

= useful heat energy (Btu)

DHW

= energy consumed by the HPWH (Wh).

The water-heating energy Q

DHW

was calculated by the data logger every 5 seconds using measured

data. These energy values were summed and logged at 15-minute intervals.

( )

inoutpDHW

TTVCQ −=



(2)

where

out

= outlet water temperature (°F)

= inlet water temperature (°F)



= hot water volumetric flow rate (ft

/h)

= specific heat of water (Btu/lbm∙°F)

= density of water (lbm/ft

4 Performance Results

Measured performance, rated capacity, EF, and FHR for each monitored model are shown in Table

5. All systems were set to hybrid mode, but set point temperatures varied. The electric resistance

percentage represents the fraction of electricity consumed by the electric resistance elements (rather

than the heat pump, controls, or peripherals). This is a small sample and many variables affect water

heater performance; however, the values do provide insight into some differences between the units.

Table 5. Summary Statics of Performance by Model

Model

Capacity

(gal)

FHR

Measured

Average

COP

Range

% Electric

Resistance

1.82

/1.61

1–2.1

32.7%

/44%

A.O. Smith

60/80

68/84

2.12

2.1–2.1

5.6%

Stiebel Eltron

78.6

2.32

2–2.6

5.5%

FHR is measured in gal/hour.

Average COP calculation for the GE units does not include Site 5 (cold air = high electric resistance use)

Average COPs over the entire monitoring period were influenced heavily by storage volumes, set

point temperature, and the ability to meet high demand over short periods of time. While the A.O.

Smith and Stiebel Eltron units used the heat pump to provide the vast majority of the load

(approximately 95% of the total electricity was consumed by the heat pump), the electric resistance

elements in the GE units consumed almost one-third of the measured electricity (excluding Site 5,

where low ambient temperatures forced the unit into resistance mode most of the time). The A.O.

Smith and Stiebel Eltron models benefit from larger storage tanks, and the Stiebel Eltron model

benefits from a factory-set set point temperature of 140°F, which increases the availability of hot

water. Increased hot water availability can increase COP by minimizing electric resistance use.

More detailed summary statistics for each site, which include operational conditions and efficiency

values, are shown in Table 6. Although the two A.O. Smith units had remarkably similar COPs, the

COPs of the Stiebel Eltron and GE units varied significantly between sites. The difference between

the COPs of the Stiebel Eltron sites is largely attributable to the large difference between the

average daily hot water draws at the two sites. The residents at Site 2 used an average of 73 GPD

with a COP of 2.6; the residents at Site 10 used an average of 41 GPD with a COP of 2. Larger hot

water draw volumes dilute the impact of tank thermal losses and elevate the COP of the unit—as

long as larger draws do not increase use of resistance heating. The differences between the

measured COPs at the sites with GE units were significantly larger than those of the other units.

With a smaller tank, the GE model seems to require more electric resistance heating to meet the

demand.

Analysis of the data collected during the year of monitoring uncovered key variables that affect

HPWH performance and the differences between the operations of the different units. Across all

models, ambient temperature, the volume of hot water draws, and the pattern of the hot water draws

were the most important variables that affected water heater performance. These variables,

particularly the effect of hot water use and electric resistance element operation, can have different

impacts on different HPWH models.

Table 6. Summary Table of Performance by Site

Site

HPWH

Model

Adults +

Children

Water Heater

Set Point Temp

(°F)

Days Monitored

Avg. Daily Hot

Water Use (gal)

Avg. Cold Water

Temp. (°F)

Avg. Hot Water

Temp. (°F)

% of Electricity

Consumption

from Electric

Resistance

Avg. Air Temp.

(°F)

Avg. Relative

Humidity (%)

Avg. Wet Bulb

(°F)

Total COP

A.O.

Smith-80

2 + 1 120 454 44 54 119 0% 59 47% 49 2.1

Stiebel

Eltron

5 + 0 140 438 73 57 136 8% 71 45% 58 2.6

2 + 1

125

469

121

48%

38%

1.8

A.O.

Smith-60

3 + 0 120 445 45 53 119 11% 63 56% 54 2.1

2 + 0

129

460

127

78%

62%

2 + 0

122

475

118

55%

2.1

2 + 0

125

450

123

11%

49%

1.8

2 + 1

125

430

122

15%

44%

2.1

2 + 2

120

468

122

22%

48%

Stiebel

Eltron

2 + 0 140 424 41 57 138 2% 68 55% 58 2

2 + 3

140

459

136

58%

34%

1.5

2 + 1

130

492

128

29%

46%

1.9

2 + 0

130

388

126

15%

57%

1.4

GE 2 + 0 120 433 32 53 119 15% 62 52% 52 1.9

Average of daily averages

Electric resistance percentage = % of total electricity kilowatt-hours that was used by electric resistance, NOT the

thermal energy fraction provided by electric resistance.

4.1 Impact of Water Use on Efficiency

As previously discussed, the smaller volume GE units used the electric resistance elements to

provide a much larger percentage of the needed energy than the other models. Figure 4 shows seven

days of operation for the GE unit at Site 9. These days are representative of typical operation across

many sites. During day 1 and days 3–6, the water draws from the tank were distributed throughout

the day and/or were relatively small. The heat pump could thus meet recovery needs for those days.

During 11/22 and 11/27, the water draws were more concentrated, and the electric resistance

element was needed to provide additional hot water.

Figure 4 shows how large and concentrated draws can reduce the efficiency of the unit; however,

low water use can also result in low COPs (Figure 5). In this figure, although the heat pump was

used to meet the entire hot water demand, the low load and relatively high standby losses resulted in

low daily COPs.

Figure 4. Ideal operation of the GE HPWH in hybrid mode (Site 9)

Figure 5. Low water use can reduce the overall benefit of an HPWH

when the cost benefit is assessed.

Scatter plots of the daily hot water use

versus the daily COP are shown in Figure 6. Each daily

observation is also color coded by the electric resistance fraction, where blue is zero electric

For this and similar figures, the vertical axes show discrete values for 15-minute intervals.

Daily values are misleading because the net energy content of the water in the tank can change throughout the day.

Thus, heating events with zero net tank energy change are identified from the 15-minute raw data. All heating events are

normalized to daily values and displayed graphically. See Appendix D for more information.

200

400

600

800

1,000

1,200

11/21 11/22

11/23 11/24 11/25

11/26 11/27 11/28

Energy Used (Wh)

HW Used (Gallons)

HW Use Heat Pump Electric Resistance

44 gals

2.1 COP

57 gals

1.4 COP

76 gals

2.6 COP

92 gals

2.4 COP

40 gals

2.4 COP

53 gals

2.4 COP

80 gals

1.5 COP

120

160

200

240

8/15 8/16 8/17 8/18 8/19 8/20 8/21 8/22

Energy Used (Wh)

HW Used (Gallons)

HW Use Heat Pump Electric Resistance

36 gals

1.7 COP

20 gals

0.9 COP

39 gals

1.9 COP

47 gals

1.4 COP

11 gals

1.3 COP

20 gals

1.5 COP

35 gals

1.4 COP

resistance use, and red is when the electric elements are used to reheat the entire recovery load. The

GE unit has the most observations; therefore, the operational patterns can be most directly observed

in the scatter plot. For the GE unit, the data could be interpreted as lying between two exponential

curves, where the upper curve has no electric resistance use and the lower curve is entirely electric

resistance operation. The scatter plot of the GE unit operation also shows that the electric resistance

fraction has a strong correlation with the hot water use. Days of lower hot water demand are far less

likely to have instances of electric resistance heating (e.g., less than 30 GPD). On the other hand,

days with high hot water demand are far more likely to have instances of electric resistance heating.

Although the scatter plots for the Stiebel Eltron and A.O. Smith units have fewer apparent trends—

because fewer sites were monitored—the scatter plots appear to agree with the trend observed for

the GE units, with one primary distinction. The GE unit shows two distinct operating regions that

correspond to large and small electric resistance loads, and the A.O. Smith and Stiebel Eltron units

operate entirely in the low electric resistance region. This is most likely a product of the larger tank

volumes and the control logic that allows the heat pump to reengage, in the case of A.O. Smith, or

operate simultaneously, in the case of Stiebel Eltron.

4.2 Impact of Air Temperature on Efficiency

Although hot water demand is a primary driver of efficiency, other factors still play a prominent

role (Figure 6). The efficiency of the heat pump is primarily a function of the ambient temperature

in the space. Two sites with cold and hot operating environments are compared to give a sense of

how ambient temperature affects overall efficiency. At Site 5, the HPWH was installed in a cold

Figure 6. Scatter plots of daily hot water use versus COP color coded by electric resistance fraction

basement with ambient air temperatures dropping lower than 50°F from December through April.

These temperatures are close to the cut-off temperature of the heat pump in the GE unit. As a result,

the HPWH at Site 5 operated like a traditional ERWH, as shown in Figure 7, but with an added

parasitic of the heat pump cycles on and off at the cut-off temperature of the unit.

Figure 7. Because the Site 5 basement was too cold, this GE HPWH

switched primarily to electric resistance mode during the winter months.

At the other end of the spectrum, Site 11 experienced higher COPs during the winter months,

because the HPWH was placed in a boiler room (Figure 9) with ambient temperatures around 80°F

(Figure 8). At the beginning of January, the unit was able to supply hot water in large quantities

with high COPs higher than the rated EF of 2.35 (Figure 10). On 2 days during this week of data,

high water use concentrations resulted in electric resistance operation to meet demand. Furthermore,

some periods had low hot water draws. As discussed in Section 4.2, lower draw quantities can

decrease overall COP, because standby losses account for a larger percentage of the recovery load.

The high ambient temperatures boost the efficiency of the heat pump and increase the

corresponding overall COP of the unit during these periods.

200

400

600

800

1,000

1,200

1,400

1/22 1/23 1/24 1/25 1/26 1/27

Energy Used (Wh)

HW Used (Gallons)

HW Use Heat Pump Electric Resistance

72 gals

0.7 COP

75 gals

0.7 COP

80 gals

0.7 COP

58 gals

0.7 COP

96 gals

0.8 COP

Figure 8. Ambient temperature at Site 11

Figure 9. This HPWH benefited from

the waste heat of the boiler used for

space heating.

Figure 10. Typical winter performance of the GE HPWH in the Site 11 warm mechanical room

To demonstrate the relationship between ambient temperature and efficiency, daily scatter plots

similar to those in Figure 7, but color coded by average ambient temperature, are shown in Figure

11. Periods with electric resistance fractions greater than 0.04 were excluded, as were data from

Sites 5 and 13, which had lower heat pump COPs than the other units. The scatter plots show a

strong correlation between ambient air temperature and efficiency. Everything else being equal,

higher ambient temperatures improve efficiency. At a 50 GPD recovery load increasing the ambient

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

Relative Humidity

Temperature [°F]

HPWH Inlet Air Temp (°F) HPWH Inlet Air RH (%)

200

400

600

800

1,000

1,200

1,400

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

Energy Used (Wh)

HW Used (Gallons)

HW Use Heat Pump Electric Resistance

56 gals

3.1 COP

135 gals

1.3 COP

71 gals

2.6 COP

195 gals

1.2 COP

78 gals

2.6 COP

71 gals

2.7 COP

87 gals

2.8 COP

temperature from 50°F to 80°F can increase the COP of the GE and A.O. Smith units from around 2

to nearly 3.

Figure 11. Scatter plots of COPs versus daily hot water use color coded by ambient temperature.

Periods with electric resistance fractions greater than 0.04 and Sites 5 and 13 were excluded.

4.3 Delivered Water Temperature

The efficiency of water heating is important from a performance standpoint; however, the

temperature of delivered water from the HPWH is also important for customer satisfaction.

Surprisingly, graphing several days of high water use at a site with a Stiebel Eltron HPWH revealed

that this unit was able to meet very large hot water demands at surprisingly high COPs, as shown in

Figure 12. While the electric resistance element was needed to supply additional heating, the unit

performed with COPs higher than 2.6 for the 4 days of high hot water use. The Stiebel Eltron unit is

distinguished from the other two models by its ability to simultaneously use the electric resistance

and heat pump heating elements to provide heating. Furthermore, the water heater was able to

deliver water temperatures in excess of 115°F during all 4 days of high hot water demand (Figure

13). The higher COPs for the Stiebel Eltron unit are probably a result of the large tank size and the

ability of the unit to only heat water for the load directly when the heat pump cannot meet the load

rather than heating the entire tank.

To determine the ability of the HPWHs to deliver water at an acceptable delivery temperature,

normalized histograms of maximum delivered water temperature for 15-minute periods with draws

larger than 1 gal were plotted for each site (Figure 14). The red and green vertical lines represent the

set point temperature and minimum acceptable delivery temperature of 110°F (Hendron and

Engebrecht 2010), respectively. The percentage of occurrences lower than the minimum acceptable

delivery temperature is shown in the upper left corner of each plot. The normalized histograms

show that all three units deliver water hotter than 110°F at a variety of set point temperatures.

However, many units did not maintain set point for all draw occurrences, and temperatures dropped

more than 10°F lower than set point for a significant fraction of the 15-minute periods.

Unfortunately, the aforementioned method of determining whether the HPWHs can maintain the

delivered water at a minimum acceptable temperature has flaws related to the 15-minute recording

interval used in this study. Because the average delivered water temperature includes water inside

the pipe before the draw, the maximum delivered water temperature for the interval must be used.

Thus, this analysis looks only at the delivery temperature at the beginning of the draw or at the

beginning of a 15-minute interval. If the water heater runs out of hot water, the user may end the

draw and the run-out may not be recorded using this method. However, the very small percentage of

run-outs recorded at the 110°F minimum acceptable delivery criteria means that delivery

temperature is probably not dropping significantly lower than usable temperatures.

Figure 12. This Stiebel Eltron HPWH was able to maintain high COPs during high water demand.

100

200

300

400

500

600

11/22 11/23 11/24 11/25 11/26 11/27 11/28 11/29

Energy Used (Wh)

HW Used (Gallons)

HW Use Heat Pump Electric Resistance

66 gals

2.6 COP

120 gals

3.1 COP

275 gals

2.6 COP

196 gals

2.7 COP

240 gals

3.0 COP

178 gals

2.6 COP

48 gals

1.7 COP

Figure 13. Even with high water demand, the hot water temperature dropped only as low as 118°F.

105

115

125

135

145

155

11/22 11/23

11/24 11/25 11/26 11/27 11/28

11/29

Hot Water Temp (°F)

HW Used (Gallons)

HW Use Delivered HW Temp

setpoint

Figure 14. Normalized histograms of delivered water temperature. Red line indicates set point temperature. Green line indicates

minimum delivery temperature (110°F). The percentage of occurrences with water temperatures lower than 110°F is shown.

105

125

145

0.5

Site 1

A.O. Smith

Probability Density

105

125

145

0.5

Site 2

Stiebel Eltron

105

125

145

0.5

Site 3

General Electric

105

125

145

0.5

Site 4

A.O. Smith

105

125

145

0.5

Site 5

General Electric

Probability Density

105

125

145

0.5

Site 6

General Electric

105

125

145

0.5

Site 7

General Electric

105

125

145

0.5

Site 8

General Electric

105

125

145

0.5

Site 9

General Electric

105

125

145

0.5

Site 10

Stiebel Eltron

Delivered Water

Temperature (°F)

Probability Density

105

125

145

0.5

Site 11

General Electric

Delivered Water

Temperature (°F)

105

125

145

0.5

Site 12

General Electric

Delivered Water

Temperature (°F)

105

125

145

0.5

Site 13

General Electric

Delivered Water

Temperature (°F)

105

125

145

0.5

Site 14

General Electric

Delivered Water

Temperature (°F)

Histogram

Setpoint Temp.

Min Delivery Temp.

0.13%

0.16%

1.44% 0.32%

1.65% 1.63% 0.85% 1.92% 1.7%

0% 0.84% 0.51% 0.32% 0.86%

4.4 Calculating Standby Losses

Tank losses to the ambient are important contributors to tank inefficiency; indeed, for ERWHs,

tank losses represent the entirety of water heat site energy inefficiencies. Unfortunately, tank

losses from HPWHs are hard to quantify without precise laboratory equipment. The data from

this study, in conjunction with laboratory HPWHs tests conducted by Sparn et al. (2011), can be

used to estimate the tank heat loss coefficient. This analysis was possible for the GE unit only,

because the Stiebel Eltron and A.O. Smith units did not provide enough data points to perform

this analysis.

The following analysis differs from the methods outlined in the U.S. Department of Energy’s test

procedure (DOE 1998) and the National Renewable Energy Laboratory’s field-testing protocol

(NREL 2012c). These methods measure the change in the average tank temperature during a

period of no draws. The National Renewable Energy Laboratory specifies that this period should

begin 10 minutes after the heat pump has shut off, because the condensing coil still gives off heat

immediately after the reheat period.

Because internal tank temperatures were not measured in this study, a different procedure was

used to measure standby losses. The basic procedure identified idle periods during which the

tank begins fully heated, experiences no water draws, and is finally reheated. The tank losses

during that period are simply the ratio of the thermal energy input to the product of the time

elapsed since the last reheat and the temperature difference between the tank and the ambient air.

( )

amb

tank

standby

−

∆

(3)

where

UA = overall heat transfer coefficient of the tank (Btu/h-°F)

standby

= thermal energy to recover tank to set point (Btu)

tank

= average tank temperature (°F)

amb

= ambient temperature (°F)

Δt = time elapsed since the last reheat (h).

The average tank temperature was assumed be 2.5°F lower than the temperature set point, which

is equivalent to assuming an average tank temperature dead band of 5°F for recovery from an

idle period. For example, if the tank set point were 135°F, the tank would presumably start the

period at 135°F, drop to 130°F due to standby losses, and finally return to 135°F through the

recovery process. The average temperature would be 132.5°F, which is 2.5°F lower than the set

point temperature. This dead band temperature is consistent with graphs of the EF testing of the

GE unit by PG&E (2010) and other studies.

Idle periods followed by heat pump operation with no hot water demand were identified from the

15-minute data to measure the electric power necessary to recover from standby losses. These

periods satisfied the following conditions:

• No water was drawn during the period.

• The period began and ended with the termination of heat pump operation to ensure

similar stratification.

• The overall period exceeded 3 hours.

For ERWHs the electricity used is directly converted to thermal energy; for HPWHs, however,

the electricity input does not directly correspond to the thermal energy input into the tank. Thus,

laboratory performance maps discussed by Sparn et al. (2011) were used to convert electricity

energy used by the heat pump to the thermal energy delivered to the storage tank. The equation

for the efficiency of the GE heat pump is:

( )

tankwbtanktankwbwbratedhphp

TTCTCTCTCTCCCOPCOP

321,

4 +++++=

(4)

where

COP

= efficiency of the heat pump (Btu/Btu)

COP

hp,rated

= rated COP

at 57°F wet bulb and 120°F tank temperature (2.76)

tank

= average tank temperature (°F)

= ambient wet bulb temperature (°F)

= 1.192E+00

= 4.247E-02

= –3.795E-04

= –1.110E-02

= –9.400E-07

= –2.657E-04.

The tank lost during the standby period is

( )

hphp

resstandby

COPE

ECQ +

(5)

where

standby

= thermal energy to recover tank to set point (Btu)

COP

= efficiency of the heat pump (kWh/kWh)

res

= electricity used by the resistance elements (kWh)

= electricity used by the heat pump (kWh)

C = conversion factor for kWh to Btu (3.412 Btu/kWh).

The estimated time of electricity cutoff was determined by examining the previous data logging

period to correct for the error inherent in the 15-minute data logging interval. If the electricity

draw for the last interval i before cutoff was smaller than the previous interval i-1, the runtime

fraction for the cutoff interval was set as the ratio of the electricity draws, as shown in Eq. (6).











≥

−

for1

for

(6)

Robust regression

analysis was used to estimate the thermal heat transfer coefficient of the

tank:

(

)

amb

tankstandby

TTt

Q −

∆=

(7)

Regressions using Eq. (7) were performed with and without a constant (i.e., nonzero intercept).

Although the idle periods were chosen such that the water heater tank could be assumed to have

no net energy change from the beginning of the period to the end of the period, the inclusion of

the constant in one of the regressions is used to test whether this assumption is valid. Regression

results with and without a constant are shown in Table 7 and Table 8, respectively, and Figure 15

shows the two regression lines plotted against the observed data. Although the intercept estimate

is significant at the 95% confidence level and the tank loss coefficient is similar to that of other

studies, the model without the constant is arguably preferable.

Table 7. Tank Heat Loss Robust Regression Results with Constant

(number of observances = 113, degrees of freedom = 111)

Coefficient Estimate

95% Confidence

Interval

Statistic

Value

Intercept

52.821

(4.2826,101.3604)

2.1564

0.033211

Overall Heat Transfer Coefficient

3.5008

(2.7034, 4.2982)

8.6994

3.4555E-14

Table 8. Tank Heat Loss Robust Regression Results without Constant

(number of observances = 113, degrees of freedom = 112)

Coefficient Estimate

95% Confidence

Interval

Statistic

Value

Overall Heat Transfer Coefficient

4.369

(4.3169,4.4211)

166.15

6.907E-136

Robust regression is similar to a simple linear regression in this case, except points with high leverage are given

less weight in the regression analysis (Chatterjee and Price 1991). Two points (see Figure 15), which appeared to be

outliers, did not qualify for exclusion under the leverage criteria of 2/N (Chatterjee and Price 1991)—where N is the

number of observations—but did have a Cook’s distance greater than 4/N (Bollen and Jackman 1985), which

suggests that these have a large impact on the regression results. Robust regression allows these points to have lower

weights.

Figure 15. Regression results and data points for GE standby losses. Temperature difference is

between the tank average temperature and ambient air.

Because the estimate of the heat loss coefficient is the desired result from the regression, the

model with the more confined confidence interval for the slope coefficient is preferred. By

including the constant in the model, the error of the heat loss coefficient estimate at the 95%

confidence level is ±22.8%, meaning that calculations using this coefficient have an acceptably

higher error. The corresponding error for the model without the constant is only ±1.2%, which is

far more acceptable for energy balance calculations. One could also argue that a higher

confidence level should be used in this analysis, because the model represents a physical system

with a well-known response and should follow a well-described deterministic phenomenon. The

heat transfer coefficient for the model without the constant is 4.312 Btu/h-°F.

4.5 Impact of Confined Spaces

Because HPWHs remove heat from the ambient air, a sufficient volume of available air is needed

to ensure proper performance of the unit. Several laboratory studies have attempted to quantify

the impact of installing HPWHs in spaces without proper air volume. All these tests were

performed on the GE GeoSpring. Advanced Energy compared two HPWHs, one placed in a

220.5-ft

uninsulated space and another subjected to the default operating conditions. GE

recommends a volume of 750 ft

(GE 2010). The efficiency of the unit in the confined space was

10% lower than the control unit (Fitzpatrick and Murray 2011). The Florida Solar Energy Center

also tested an HPWH in a confined space (92.5 ft

) and saw a reduction in efficiency of 13%

(Colon 2012). The National Renewable Energy Laboratory’s laboratory tests included blocking

the airflow of the unit with tape. Blocking the airflow by 33% and 67% had only a minor impact

on the performance of the unit, most of which was attributable to the increased fan energy use

(Sparn et al. 2011). This suggests that the volume of air available to the unit is more important

than constriction of airflow.

40 45 50

55 60 65 70

100

200

300

400

500

600

700

800

Temperature Difference (°F)

Thermal Losses (BTU/hr)

Observances

Without Constant

With Constant

The HPWH at Site 13 was installed in a confined space with approximately 400 ft

of available

air. The heating event shown in Figure 16 resulted in a measured COP of 2.02; the resulting

performance drop was approximately 15.8% lower than an approximate expected COP of 2.4 at a

set point temperature of 130°F. The performance reduction is similar to those found by

Fitzpatrick and Murray (2011) and Colon (2012).

Figure 16. Performance of the HPWH at Site 13 in a confined space

4.6 Impact on Space-Conditioning Systems

Because HPWHs move heat from the air surrounding the water heater into the storage tank,

HPWHs have an impact on space-conditioning loads. HPWHs in conditioned space have a very

direct effect on space-conditioning loads. In the summer, HPWHs reduce the cooling load of the

building; conversely, the heating load increases in the winter. The combined impact of the

HPWH energy consumption and its impact on the space conditioning systems rely heavily on

climate, home configuration, HPWH location, and the space-conditioning systems used. For

more information about these impacts see the modeling results described by Hudon et al. (2012).

A less complex analysis was also presented by Shapiro, Puttagunta, and Owens (2012).

For HPWHs installed in unconditioned or “semiconditioned” spaces such as basements, the

space-conditioning impacts are harder to determine. These spaces act as buffers, so the heat

transferred from the space into the storage tank is not necessarily transferred from the

conditioned spaces. Air infiltration into the buffer space, ground coupling, solar gains, and heat

transfer from mechanical equipment can all affect heat transfer between the buffer space and

conditioned space. The transient effects of these heat transfer processes mean that the HPWH

space-conditioning impacts are potentially reduced.

15:00 18:00 21:00 00:00

Hot Water Flow Rate (gpm)

Site13; COP = 2.02; T = 69.6; RH = 55%; T

mains

= 57.7; T

set

= 130;

15:00 18:00 21:00 00:00

7.5

22.5

Cummulative

Volume (gal)

15:00 18:00 21:00 00:00

200

400

600

800

1000

1200

Electric Energy Used (Wh)

Flow Rate

Heat Pump

Lower Element

Upper Element

Because these transient heat transfer interactions cannot be directly addressed for the HPWHs

installed in unconditioned basements in this study, the space-conditioning impacts are addressed

as the maximum potential impact. The impact on the space-conditioning equipment is defined as

(NREL 2012c):

dhwinputspacenet,

QQQ −=

(8)

where

net,space

= maximum potential thermal impact on space (Btu)

input

= electric energy used by the water heater (Btu)

dhw

= thermal energy removed from storage tank (Btu).

The maximum potential thermal impact on the space-conditioning equipment on the space is

negative when thermal energy is being removed from the space and positive when the energy is

being added to the space. In the case of standard water heaters, Q

net,space

is positive, but for

HPWHs the net energy to the space is negative.

Although the HPWHs remove energy from the space, the impact on the space-conditioning loads

depends on the season. For this analysis, the cooling season for these sites was considered to be

June through September, and all other periods were considered the heating season. Box plots of

the hourly load and seasonal space-conditioning impacts are shown in Figure 17 (negative values

reduce the space-conditioning loads and positive values increase the space-conditioning loads).

Box plots display the range of data for each of the categories. In the plots, the boxes represent the

inner two quartiles—also called the inner quartile range—where the lower quartile is the bottom

of the box and the upper quartile is the top of the box. The red line in the middle of the box

represents the median. At the top and bottom of the boxes are whiskers that represent the

smallest and largest observations within the 1.5 inner quartile range. Outliers are represented by

a red plus, and the mean is represented by a black plus.

Figure 17. Box plots of maximum potential space-conditioning impacts for monitored HPWHs

-1000

-500

500

1000

1500

Cooling Heating

Mean Hourly Change in Load (Btu/hr)

-2

Cooling Heating

Seasonal Change in Load (MMBtu)

As shown in the box plots, the magnitude of the mean hourly heating impact is slightly higher

than that of the mean hourly cooling impact. Similarly, the magnitude of total heating seasonal

impact is nearly twice that of the total cooling seasonal impact. These box plots show that the

homes in this study, which are all in northern climates, had maximum potential seasonal cooling

load reductions of approximately 1.5 MBtu and heating load increases of approximately 3 MBtu.

5 Energy and Cost Analysis

Energy and cost savings potentials of HPWHs were investigated against eight alternative storage

tank water heaters: electric resistance; standard natural gas, fuel oil, and propane; premium

natural gas, fuel oil, propane; and condensing natural gas. Tankless water heaters were not

considered, because a more complicated modeling procedure to account for the transient

response of these water heaters would be necessary. The assumed EFs, recovery efficiencies, and

pilot light energy are listed in Table 9, where all standard and premium fossil-fuel-based systems

are assumed to have the same operating characteristics except for the energy source used to

provide the heating energy.

Table 9. Alternative Water Heaters

Water Heater Fuel Types EF

Recovery

Efficiency

Pilot Light Energy

(Btu/h)

Electric

Resistance

Electricity 0.92 100% 0

Standard

Fossil Fuel

Natural gas, fuel

oil, and propane

0.59 78% 400

Premium

Fossil Fuel

Natural gas, fuel

oil, and propane

0.67 82% 400

Condensing

Fossil Fuel

Natural gas 0.83

95%

WATSMPL default (EPRI 2000)

Efficiencies from laboratory testing by PG&E (2008)

The alternative water heaters were simulated using the equations underlying the WATSMPL

simplified water heater software (EPRI 2000). The daily operating conditions at each site—set

point temperature, mains temperature, volume of water drawn, and ambient temperature—were

simulated for each using the equations in Appendix B. The monitored (in the case of the

HPWHs) and simulated energy consumption (in the case of the alternative water heaters) were

normalized to yearly energy consumption. The normalization procedure accounts for the fact that

the monitoring period at each site was different and some overlap occurred in the days that were

monitored.

Although energy use is usually measured in site energy—which is the energy used at the home

and is typically measured at a utility meter in kilowatt-hours (electricity)—therms (natural gas),

or gallons (fuel oil or propane), a better metric for measured energy use is source energy, which

is the sum of energy used at the home and the energy lost to extraction, conversion, or

transmission. Site energy is easily converted to source energy using site-to-source ratios (Deru

and Torcellini 2007), which are defined to the fuels studied in Table 10. Energy use is thus

reported in source energy and cost to consumers. The site energy consumption for each water

heater-site combination was converted to source energy and cost using the values listed in Table

10.

Table 10. Fuel Prices and Characteristics for Energy and Cost Analysis

Energy Source

Site-to-Source Ratio

Cost

Energy Content

Electricity

3.365

$0.1172/kWh

3,412 Btu/kWh

Natural Gas

1.092

$1.10/therm

100,000 Btu/therm

Fuel Oil

1.158

$3.93/gal

140,000 Btu/gal

Propane

1.151

$2.85/gal

91,600 Btu/gal

Deru and Torcellini (2007)

Prices obtained from EIA (2012b, 2012c, 2012d). Electricity and natural gas prices represent average national

prices for 2011. Fuel oil and propane prices are the average national prices for October 2011 through March 2012.

Thumann and Mehta (1991)

To compare the installed HPWHs and the simulated alternative water heaters, box plots of

annual source energy consumption and annual operating costs are shown in Figure 18 and Figure

19 (data tables provided in Appendix C). Figure 18 shows that HPWHs have significantly lower

operating costs than electric resistance, oil, and propane, and comparable operating costs to

natural gas systems. Furthermore, HPWHs have the lowest mean and median operating costs of

all water-heating systems except condensing natural gas. In terms of source energy consumption,

Figure 19 shows that HPWHs have the second-lowest source energy consumption of all water-

heating products (a natural gas condensing tank unit is the lowest). Unlike ERWHs, which have

the highest source energy consumption of all products by far, HPWHs can counteract the source

energy penalty of using electricity to provide water heating. The outlier for the HPWHs in both

plots is Site 5, which was operating in electric resistance mode throughout the winter months and

therefore has operating costs similar to the ERWH category.

Figure 18. Annual operating cost of monitored HPWHs and alternative water heaters

$100

$200

$300

$400

$500

$600

$700

$800

$900

Heat

Pump

Electric

Resistance

Gas

Standard

Gas

Premium

Gas

Condensing

Oil

Standard

Oil

Premium

Propane

Standard

Propane

Premium

Annual Operating Cost ($/yr)

Figure 19. Annual source energy of monitored HPWHs and alternative water heaters

Although HPWHs have the second-lowest operating costs and source energy use of the water

heaters considered in this study, installation costs have an impact on the overall financial

outcome of HPWHs versus alternative water heaters. The cost analysis was performed as

described by Polly et al. (2011). The assumptions and results are listed in Table 11. The costs

were assumed to be wrapped into a 30-year mortgage at a 4.42% interest rate. These assumptions

are based on wrapping the cost of the water heaters into the cost of buying a new house or

mortgage refinance.

Table 11. Cost Analysis Assumptions

Cost Metric

Assumption

Analysis Period

30 years

Inflation Rate

Real Discount Rate

Real Fuel Escalation Rate

Mortgage Rate

4.42%

Mortgage Period

30 years

Marginal Income Tax Rate

28%

Building Energy Optimization defaults (NREL 2012b)

Average rate for 2011 (Freddie Mac 2012)

Table 12 lists the installation costs and typical lifetimes of the various water-heating products

considered. Using these values, the annualized energy-related cost of each water-heating unit

was calculated. The annualized energy-related cost represents the equivalent annual cost of the

Heat

Pump

Electric

Resistance

Gas

Standard

Gas

Premium

Gas

Condensing

Oil

Standard

Oil

Premium

Propane

Standard

Propane

Premium

Annual Source Energy (MMBtu/yr)

complex cash flow and represents an equivalent annual cost of operating the water heater in

present dollars.

Table 12. Installation Costs and Lifetimes of Water-Heating Products

Water Heater Installation Cost

Typical Lifetime

(years)

Heat Pump (80 gal)

$3,300

Heat Pump (50-60 gal)

$2,100

Electric Resistance

$590

Standard Natural Gas

$700

Premium Natural Gas

$880

Condensing Natural Gas

$4,500

Standard Fuel Oil

$820

Premium Fuel Oil

$960

Standard Propane

$890

Premium Propane

$1,400

From National Residential Efficiency Measures Database (NREL 2012a).

From American Water Heaters (2008)

Estimated lifetime based on National Residential Efficiency Measures Database values for gas storage and

condensing tankless heaters.

The installation costs for the cost analysis were derived from the National Residential Efficiency

Measures Database, but HPWH installation costs reported for this study (Table 13) were

generally in line with the costs reported in the National Residential Efficiency Measures

Database. These reports show shorter lifetimes for the HPWHs than the other water heater

technologies, which is probably caused a desire to be conservative and rate the lifetime equal to

the product warranty period. The reliability of these units is as yet undetermined, however, so the

lifetime of HPWHs may be similar to other water heaters.

Table 13. Installed HPWH Cost Estimates from Study

Small Tank

(50–60 gal)

Large Tank

(80 gal)

Unit

$1,399

$2,403

Extra Labor

$69

Condensate Pump

$154

Electric and Plumbing Permit

$100

Breaker

–

$54

Tempering Valve

–

$142

Labor

$200–$400

$400-$600

Total

$1,922–$2,122

$3,318-$3,518

A boxplot showing the spread of the annualized energy-related cost for the nine water-heating

technologies is shown in Figure 20. The HPWHs have lower annual energy-related costs than the

electric resistance, fuel oil, and propane water heaters, although the annualized energy-related

cost reduction over ERWHs is smaller than the annual operating cost savings due to the

increased first cost of HPWHs. As expected, the noncondensing natural gas water heaters are still

the lowest-cost option on a total life cycle basis. Space-conditioning interactions may change the

relative costs of HPWHs in relation to other water heaters, depending on the climate. These

interactions could not be measured in this study. Condensing tankless water heaters, which were

not investigated in this analysis, have lower operating costs than even condensing storage water

heaters, but the analysis method employed here does not account for the transient behavior of

tankless water heaters. Tankless units have other issues, so a direct comparison may not be

appropriate.

Figure 20. Annualized energy-related costs of monitored HPWHs and alternative water heaters

$200

$300

$400

$500

$600

$700

$800

$900

$1000

Heat

Pump

Electric

Resistance

Gas

Standard

Gas

Premium

Gas

Condensing

Oil

Standard

Oil

Premium

Propane

Standard

Propane

Premium

Annuallized Energy Related Cost ($/yr)

6 Homeowner Surveys

After complete monitoring of the HPWHs, residents at the 14 sites were surveyed about their

satisfaction with the units. All the homeowners were satisfied. The majority (70%) noticed

cooling and/or dehumidification. Some noted that noise was an issue (18%). Some homeowners

experienced running out of water (36%), though these were isolated incidences and were not a

significant concern. The majority (73%) noticed lower utility bills. One “No” response to

noticing utility bill savings was because the homeowner switched from oil and did not know how

to do the cost comparison. Another “No” response was that the homeowner was not sure if the

savings were a result of HPWH or lower electricity rate (the savings were indeed due to the

HPWH). Detailed survey results are shown in Table 14 and Table 15.

Table 14. Survey Results of Whether the Homeowner or a Qualified

Professional Performed any Preventive Maintenance Procedures

GE AO SE All

Y N Y N Y N Y N

Testing the temperature & pressure-relief valve?

Flushing and/or draining the tank?

Periodically inspecting and clearing the condensate strainer

and drain lines?

5 3 1 0 0 2 6 5

Visually inspecting the surrounding floor area, or the drain

pan for signs of water leakage?

3 5 1 0 2 0 6 5

Cleaning the air filter?

Cleaning the evaporator?

Checking the condition of the sacrificial anode rod?

Checking and descaling the heating elements?

Would you encourage a friend or family member to buy the

same water heater?

8 0 1 0 1 1 10 1

The ideal response is in green text and nonideal response is in red text. Questions in gray were not expected to be addressed because these units

have not been installed long enough to warranty this level of maintenance

Table 15. Survey Results of Homeowners

GE AO SE All

Y N Y N Y N Y N

Have you ever run out of hot water?

4 4 0 1 0 2 4 7

Has the water been hot enough?

Have you noticed a difference in your energy bills since the water

heater was installed?

6 2 1 0 1 1 8 3

Have you noticed the water heater cooling and/or dehumidifying

the space?

6 2 0 1 1 0 7 3

Has the water heater’s operating noise been a problem?

Has this water heater changed how you use hot water?

Are you satisfied with your HPWH’s performance?

8 0 1 0 2 0 11 0

Have you read your HPWH’s manual?

Do you know how to change the settings on your HPWH?

Have you ever changed the settings of the water heater?

Do you know what operating mode your HPWH is set to?

Has the water heater required servicing?

The ideal response is in green text, nonideal response is in red, and questions without an ideal response are in blue.

7 Recommendations

Based on the results of this report, the Consortium for Advanced Residential Buildings provides

several recommendations for successful installation of HPWHs. The efficiency of HPWHs is

profoundly affected by hot water use. When large quantities of hot water are used in clusters,

HPWHs revert to electric resistance mode, which reduces the efficiency of the unit. A

homeowner can reduce this effect by purchasing a larger HPWH, increasing the set point

temperature (only for high water draw users, low water draw users will be negatively impacted

by the higher standby loss resulting from a higher set point temperature), or changing behavior.

By increasing the size and temperature of an HPWH, more hot water can be delivered at a given

time before the resistance elements are needed. Spreading the water load over a longer period of

time may also provide similar benefits and reduce standby losses.

The location of an HPWH is also an important factor to consider during installation. HPWHs

operate at higher efficiencies when they are subjected to higher ambient temperatures. If

possible, HPWHs should not be installed in cold locations, such as garages in northern climates.

Furthermore, HPWHs require additional space to operate at peak efficiency. In this and other

studies, an efficiency reduction of more than 10% was observed when HPWHs are installed in

confined spaces. Finally, HPWHs remove heat from the surrounding air and can therefore affect

the space-conditioning loads of a building. In cold climates, this heat removal is typically

undesired, and installing an HPWH in a buffer space such as an unconditioned basement can

temper the space-conditioning affects. In hot climates, removing heat can be beneficial, and the

total building energy consumption can be reduced by placing HPWHs in the conditioned space.

Although HPWHs are a promising technology that may finally be here to stay, installers must

still address some hurdles. Successful installation requires careful consideration of clearance and

weight. With additional heating mechanisms, these HPWHs are markedly larger than ERWHs

with identical storage volumes. Because HPWHs remove humidity from the air, drain pans and

condensate pumps may be necessary to protect them and the floor from excess moisture. Finally,

the inclusion of a heat pump means that filters must be cleaned regularly and noise could be a

problem for some installations. Many of these details are discussed in more depth by Shapiro,

Puttagunta, and Owens (2012) and SWA (2012).

8 Conclusion

Though the study used a small sample set, the overall performance of these 14 HPWHs has been

enlightening and shows promise for this technology. To date, only one compressor for an HPWH

unit had to be replaced; the cause of this failure is unclear. No other major issues have been

identified about the durability and reliability of these units, but this will need to be followed up

as these systems age. This evaluation successfully answered the following research questions:

• What is the measured efficiency of an HPWH located in unfinished basements of cold-

climate homes?

Even when installed in unfinished basements of cold-climate homes, the measured

efficiency of these units was much higher than that of conventional ERWHs (Table 16).

Except for one unit that was placed in a basement with very low ambient temperatures,

the monitored units had COPs higher than 1.5, which represents more than a 50%

reduction in energy consumption compared to an ERWH. The highest COP was 2.6.

Table 16. Summary of Test Results by Model

Model

Units

Monitored

Capacity

(gal)

Measured

Average COP

COP

Range

1.82

1.5–2.1

A.O. Smith

60/80

2.12

2.1

Stiebel Eltron

2.32

2.0–2.6

Annual operating costs and source energy consumption for the monitored HPWHs were

lower than those for the alternative storage tank water heaters considered. Annualized

energy-related costs for HPWHs were slightly lower than those for ERWHs and

considerably lower than those for propane- and fuel-oil-fired systems. Natural gas storage

water heaters, however, were still the lowest-cost storage water heater on an annualized

energy-related cost basis. Tankless water heaters were not considered here because of

their more complex modeling requirements.

• What are the critical criteria that affect the installed performance of HPWHs, and how do

they impact the performance of each HPWH model?

Determining the key variables that affect performance was based on the results of the

study data.. Domestic hot water use (daily volume and draw pattern) is the primary driver

of efficiency, but ambient temperature plays a considerable role in expected efficiency.

Lower water demand reduces the overall efficiency of the water heater by increasing the

fraction of the recovery load that is standby losses. High water demand also reduces

efficiency by requiring more electric resistance use. Higher ambient temperatures

increase the efficiency of the heat pump and reduce standby losses.

Although the GE unit shows two distinct operating regions that correspond to large and

small electric resistance loads, the A.O. Smith and Stiebel Eltron units operate entirely in

the low electric resistance region. This is most likely a product of the larger tank volumes

and the control logic that allows the heat pump to reengage, in the case of A.O. Smith, or

operate simultaneously, in the case of Stiebel Eltron. The Stiebel Eltron unit’s higher

factory-set set point temperature may also play a role.

• Does each model evaluated in this study effectively deliver hot water at the set point

temperature?

Despite the slower recovery rate of heat pump compared to electric resistance elements,

the HPWHs were surprisingly able to meet very high hot water demands. The Stiebel

Eltron unit in particular was able to meet large loads around 200 GPD of hot water at

high COPs. The outlet water temperature during these periods never dropped below the

minimum accepted water temperature of 110°F.

• What are the standby losses for these HPWHs, and how do they compare to traditional

ERWHs?

HPWHs are significantly more efficient than ERWHs; however, the standby losses of

these water heaters are considerably higher than those of ERWHs. Laboratory

measurements of these units in resistance mode showed values about 10% lower than a

comparably sized ERWH. Laboratory testing and analysis in this study also showed

higher overall heat transfer coefficient values than those for ERWHs. Tank loss

coefficients calculated in this study were about 25% higher than laboratory tests, which

may reflect the real-life performance of these units better than the laboratory installations.

A possible explanation for this difference is that thermal shorts were introduced by the

wraparound heat exchanger. Improvements in tank insulation would need to be made to

reduce the thermal heat loss for these tanks.

Although the efficiency numbers of these units are impressive, this study did not

completely address the interactions between the HPWHs and the space-conditioning

systems. Because the HPWHs were placed in unconditioned basements, which act as

buffer spaces, the total impact on the space-conditioning system cannot be measured

without more information. However, the maximum potential impact of the HPWH on the

conditioning loads of the houses was approximately 3 MBtu of increased heat load and

1.5 MBtu of decreased cooling load. In other climates, the overall effect may be quite

different.

• Are homeowners satisfied with hot water delivery, efficiency, noise, and other

characteristics?

Ten of the 11 survey respondents said that they would recommend an HPWH to a friend

or family member. The one dissenting homeowner had an issue with the noise of the

HPWH, because a home office was located in the room adjoining the basement

mechanical room.

Overall, this study provides considerable data about the performance of new HPWHs in

unconditioned basements in the Northeast. The efficiencies were remarkable in these

installations, and these units show considerable promise. However, more research is still

necessary in several areas. Understanding the affects of HPWH installation in unconditioned

basements is vital for quantifying HPWH impacts on the total building energy use. Calculating

the interactions among the HPWH, conditioned space, buffer space, ground, and ambient air

temperature is not a trivial task.

Furthermore, more information is needed to predict the effect of draw profiles on the efficiency

of these HPWHs. Although the concentration of draws clearly affects the switching between the

heat pump and the electric resistance elements, a direct connection between which draw profiles

trigger resistance operating and which do not has not been drawn. Understanding this connection

is particularly vital for developing test procedures that can accurately predict the field

performance of HPWHs.

Although HPWHs are a promising technology that may finally be here to stay, installers must

still address some installation hurdles. Successful installation requires careful consideration of

clearance and weight, drain pans and condensate pumps, maintenance, and noise. Many of these

details are discussed in more depth by Shapiro, Puttagunta, and Owens (2012) and SWA (2012).

References

Abdelaziz, O., K. Wang, E.A. Vineyard, and J. Roetker, J. 2012. “Development of

Environmentally Benign Heat Pump Water Heaters for the US Market.” ACEEE Summer Study

on Energy Efficiency in Buildings, 1.1–1.12.

AIL Research. 2001. Northeast Utilities Heat Pump Water Heater Field Test: 30 Crispaire

Installations. Princeton, NJ: AIL Research Inc.

AIL Research. 2002. Nyle Heat Pump Water Heat Evaluation. Princeton, NJ: AIL Research Inc.

Amarnath, A., and J. Bush. 2012. “HPWH Field Demonstration Project: Overview and Some

Interim Results.” Presented at ACEEE Hot Water Forum. May 21–23, 2012.

American Water Heaters. 2009. Polaris

High-Efficiency Commercial Gas Water Heater.

NCGSS00308 R April 09. Johnson City, TN: American Water Heaters.

A.O. Smith. 2010. A.O. Smith—Voltex

Hybrid Electric Heat Pump Water Heater: Installation

Instructions and Use & Care Guide. 318257-000. February 2010. Ashland City, TN: A.O. Smith.

A.O. Smith. 2011. Personal Correspondence. March 11, 2011. A.O. Smith, Ashland City, TN.

Ashdown, B.G., D.J. Bjornstad, M. Boudreau, V. Lapsa, S. Schexnayder, B. Shumpert, and F.

Southworth. 2004. Heat Pump Water Heater Technology: Experiences of Residential Consumers

and Utilities (Technical Report, ORNL/TM-2004/81). Oak Ridge, TN: Oak Ridge National

Laboratory.

Baxter, V.D., and R.L. Linkous. 2004. Heat Pump Water Heater Durability Testing—Phase II

(Technical Report, ORNL/TM-2004/111). Oak Ridge, TN: Oak Ridge National Laboratory.

Bodzin, S.1997. “Air-to-Water Heat Pumps for the Home.” Home Energy. July/August 1997.

Bollen, K.A., and R.W. Jackman. 1985. “Regression Diagnostics: An Expository Treatment of

Outliers and Influential Cases.” Sociological Methods Research; 13(4): 510–542.

BLS. 2012. “CPI Inflation Calculator.” Washington, DC: U.S. Bureau of Labor Statistics.

Accessed October 31, 2012: www.bls.gov/data/inflation_calculator.htm.

Calm, J.M. 1984. Heat Pump Water Heaters. Palo Alto, CA: Electric Power Research Institute.

Changing Times. 1982. “Heating Water with a Heat Pump.” The Kiplinger Magazine: Changing

Times. August 1982; pp. 18–19.

Chatterjee, S., and B. Price. 1991. Regression Analysis by Example. 2

Edition. New York: John

Wiley & Sons.

Colon, C. 2012. Side-by-Side Testing of Water Heating Systems: Results from the 2010-2011

Evaluation. FSEC-RR-386-12. Cocoa, FL: Florida Solar Energy Center.

DOE. 1998. “Energy Conservation Program for Consumer Products: Test Procedure for Water

Heaters.” Federal Register; 10 CFR Part 430; 63(90): 25996–26016. Washington, DC: U.S.

Department of Energy.

DOE. 2010. “Energy Conservation Program: Energy Conservation Standards for Residential

Water Heaters, Direct Heating Equipment, and Pool Heaters.” Federal Register; 10 CFR Part

430; 75(73): 20114–20236. Washington, DC: U.S. Department of Energy.

Deru, M., and P. Torcellini. 2007. Source Energy and Emission Factors for Energy Use in

Buildings (Technical Report, NREL/TP-550-38617). Golden, CO: National Renewable Energy

Laboratory.

EIA. 1996. The Changing Structure of the Electric Power Industry: An Update. DOE/EIA-

0562(96). Washington, DC: U.S. Energy Information Administration.

EIA. 2005. “2005 Residential Energy Consumption Survey.” Washington, DC: U.S. Energy

Information Administration. Accessed October 31,

2012: www.eia.gov/consumption/residential/data/2005/.

EIA. 2009. “2009 Residential Energy Consumption Survey.” Washington, DC: U.S. Energy

Information Administration. Accessed October 31,

2012: www.eia.gov/consumption/residential/data/2009/

EIA. 2012a. Annual Energy Outlook 2012: with Projections to 2035. DOE/EIA-0383.

Washington, DC: U.S. Energy Information Administration.

___ 2012b. “Electricity Data Browser.” Washington, DC: U.S. Energy Information

Administration. Accessed October 31, 2012: www.eia.gov/beta/enerdat/.

___ 2012c. “U.S. Natural Gas Prices.” Washington, DC: U.S. Energy Information

Administration. Accessed October 31,

2012: www.eia.gov/dnav/ng/ng_pri_sum_dcu_nus_m.htm

___ 2012d. “U.S. Weekly Heating Oil and Propane Prices.” Washington, DC: U.S. Energy

Information Administration. Accessed October 31,

2012: www.eia.gov/dnav/pet/pet_pri_wfr_dcus_nus_m.htm.

___ 2012e. “Custom Table Builder” Washington, DC: U.S. Energy Information Administration.

Accessed November 2, 2012: www.eia.gov/forecasts/steo/query/.

Environmental Building News. 2005. “WatterSaver Water Heater Halted.” Environmental

Building News. 14(11).

EPA. 2012. “Residential Water Heaters Key Product Criteria.” Washington, DC: U.S.

Environmental Protection Agency. Accessed October 31,

2012: www.energystar.gov/index.cfm?c=water_heat.pr_crit_water_heaters.

EPRI. 1995. “Energy Consumption and Operating Cost Comparison for Water Heaters.” Electric

Water Heating News. 7(4). Palo Alto, CA: Electric Power Research Institute.

EPRI. 2000. WATSMPL 2.0 User’s Manual: Simplified Energy and Operating Cost Analysis of

Storage Water Heaters. Palo Alto, CA: Electric Power Research Institute.

Fitzpatrick, S., and M. Murray. 2011. GE Heat Pump Water Heater Report. Raleigh, NC:

Advanced Energy.

Franco, V., A. Lekov, S. Meyers, and V. Letschert. 2010. “Heat Pump Water Heaters and

America Homes: A Good Fit?” Proceedings Summer Study on Energy Efficiency in Buildings

(9); American Council for an Energy-Efficient Economy; pp. 68–79.

Freddie Mac. 2012. “30-Year Fixed-Rate Mortgages Since 1971.” McLean, VA: Freddie Mac.

Accessed October 31, 2012: www.freddiemac.com/pmms/pmms30.htm.

GE. 2009. Hybrid Residential Water Heaters: Owner’s Manual & Installation Instructions. 49-

50254 10-09 JR. Louisville, KY: General Electric Appliances.

Hashimoto, K. 2006. “Technology and Market Development of CO

Heat Pump Water Heaters

(ECO CUTE) in Japan.” IEA Heat Pump Centre Newsletter; 24(3): 12–16.

Hendron, R., and C. Engebrecht. 2010. Building America House Simulation Protocols.

DOE/GO-102010-3141. Golden, CO: National Renewable Energy Laboratory.

Hu, S., and R. Davis. 2011. “PG&E Field Study on Heat Pump Water Heaters.” Presented at

ACEEE Hot Water Forum. May 10, 2011.

Hudon, K., B. Sparn, D. Christensen, and J. Maguire. 2012. “Heat Pump Water Heater

Technology Assessment Based on Laboratory Research and Energy Simulation Models.”

Preprint. NREL/CP-550-51433. Golden, CO: National Renewable Energy Laboratory.

Maynard, M. 2011. “Advances in Water Heating Technology.” AHRItrends. Fall 2011.

Megliola, M. 2011. Personal Correspondence. March 11, 2011. Stiebel Eltron, West Hatfield,

MA.

Murphy, R.W., and J.J. Tomlinson. 2002. Field Tests of a “Drop-in” Residential Heat Pump

Water Heater (Technical Report, ORNL/TM-2002/207). Oak Ridge, TN: Oak Ridge National

Laboratory.

NREL. 2012a. “National Residential Efficiency Measures Database.” Golden, CO: National

Renewable Energy Laboratory. Accessed October 30,

2012: www.nrel.gov/ap/retrofits/about.cfm.

____ 2012b. BEoptE+ 1.4. https://beopt.nrel.gov/home.

____ 2012c. Field Monitoring Protocol: Heat Pump Water Heaters. Golden, CO: National

Renewable Energy Laboratory (Forthcoming).

PG&E. 2008. Laboratory Testing of Residential Gas Water Heaters. ATS Report #491-08.5. San

Francisco: Pacific Gas & Electric Company.

PG&E. 2009. Energy Performance Analysis for Heat Pump Water Heaters. Application

Assessment Report #0916. San Francisco: Pacific Gas & Electric Company.

PG&E. 2010. Laboratory Evaluation and Field Testing of Residential Heat Pump Water Heaters.

ETP-10PGE1001. San Francisco: Pacific Gas & Electric Company.

Polly, B., M. Gestwick, M. Bianchi, R. Anderson, S. Horowitz, C. Christensen, and R. Judkoff.

2011. A Method for Determining Optimal Residential Energy Efficiency Retrofit Packages.

DOE/GO-102011-3261. Golden, CO: National Renewable Energy Laboratory.

Powell, E. 1980. “New Energy-Saving Heat-Pump Water Heater.” Popular Science, April 1980,

pp. 49–57.

Shapiro, C., S. Puttagunta, and D. Owens. 2012. Measure Guideline: Heat Pump Water Heaters

in New and Existing Homes (Subcontract Report, DOE/GO-102012-3456). Golden, CO:

National Renewable Energy Laboratory.

Sparn, B., K. Hudon, and D. Christensen. 2011. Laboratory Performance Evaluation of

Residential Integrated heat Pump Water Heaters (Technical Report, NREL/TP-550-52635).

Golden, CO: National Renewable Energy Laboratory.

Stiebel Eltron. 2010. Operating and Installation Manual: DHW Heat Pump Water Heater. West

Hatfield, MA: Stiebel Eltron.

SWA. 2004. Field Evaluation of WatterSaver

™

Heat Pump Water Heater. Norwalk CT: Steven

Winter Associates.

SWA. 2012. Heat Pump Water Heaters: Selection and Quality Installation Guide. Springfield,

MA: Mass Save. Accessed October 20,

2012:

www.masssave.com/~/media/Files/Residential/Information%20and%20Edu%20Docs/HP

WH_QI_Guide.ashx.

Thumann, A., and P. Mehta. 1991. Handbook of Energy Engineering. 2

Ed. Lilburn, GA:

Fairmont Press.

Tomlinson, J.J. 2002. “All Pumped Up: New Heat Pump Water Heaters Consume Half the

Energy of Conventional Electric-Resistance Ones.” Home Energy. November/December; pp. 30–

35.

Wilkes G., and F.M. Reed. U.S. Patent No. 2,095,017 (October 5, 1937).

Zogg, R. 2002. “Residential Heat Pump Water Heater Technology and Markets Seminar.”

December 4–5, 2002; Portland, OR.

Zogg, R.A., W.J. Murphy, and J. Hoyt. 2004. Design Refinement and Demonstration of a

Market-Optimized Heat-Pump Water Heater. CEC-500-04-018. Sacramento, CA: California

Energy Commission.

Appendix A: A History of Heat Pump Water Heaters in the United

States

HPWHs are not new, despite their recent reintroduction into the mainstream residential

marketplace. Since the first commercialization of HPWHs in the 1950s, HPWHs have followed a

boom-bust cycle of product development and subsequent abandonment that closely follows retail

electricity prices and demonstrates the reliability issues of past products. Even though HPWHs

failed to gain significant market traction in the past, the current confluence of regulatory,

economic, and market forces may be finally pushing HPWHs into the mainstream.

The Early Years: 1935–1970

The invention of HPWHs dates back at least to 1935, when the first HPWH was patented in the

United States (Wilkes and Reed 1937).

The HPWH described in the patent (Figure 21) looks

very similar to the current batch of HPWHs, with a supplemental electric resistance element, an

integrated storage tank, a motor-driven compressor, and an immersed coil condenser. A unique

feature of this HPWH, however, was the use of upper and lower chambers to prevent the electric

resistance element from heating the entire storage volume.

Figure 21. First patented HPWH. Image from Wilkes and Reed 1937.

The search term heat pump water heater was used in Google’s patent search, which has digitized every patent

application in the United States since 1790.

Development of the first mass-market, commercialized HPWH, however, did not begin until the

1950s. A prototype developed by the Hotpoint Company—now a division of GE—performed

well when monitored in collaboration with the Tampa Electric Company (Calm 1984).

Throughout the 1940s and 1950s electricity prices declined considerably

(EIA 1996), which

undoubtedly reduced the economic and commercial viability of the product. Development was

discontinued, and HPWHs would not return to the mainstream market for more than 20 years

(Calm 1984).

The Second Generation: 1970–1999

Interest in HPWHs subsided in the 1960s, during a period of low and falling electricity prices.

By the 1970s, however, the emerging energy crisis and a spike in retail electricity prices

rekindled interest in HPWHs as a mainstream, commercialized product. Retail electricity prices

began to rise as fossil fuel prices spiked, utility companies overbuilt capacity based on inaccurate

demand forecasts, and environmental regulation increased construction costs for capital

improvements (EIA 1996).

During the mid- to late 1970s, the National Rural Electric Cooperative Association and the U.S.

Department of Energy provided funding to develop a prototype HPWH. Through this funding,

Energy Utilization Systems manufactured 100 HPWHs—85 integrated units and 15 add-on

units—which were tested by 20 electric utilities. The add-on units were unsuccessful; however,

the integrated units showed significant promise. Average COPs were around 2, operating savings

approached 50%, the life span of the units matched similar natural gas water heaters and

ERWHs, and consumer satisfaction was high (Calm 1984).

By the 1980s, the successful pilot program had driven HPWHs into the mainstream marketplace.

Popular Science trumpeted HPWHs as a way to achieve “solar savings—without the cost”

(Powell 1980), and HPWHs were soon eligible for state tax credits. Electric utilities, such as the

Tennessee Valley Authority, pushed adoption through zero-interest loans and incentives

(Changing Times 1982). By 1984, at least 17 commercialized HPWHs were in the marketplace

under a multitude of trade names (Calm 1984).

These new units were met with enthusiasm during the early 1980s and sold more than 10,000

units per year. However, the HPWH market soon collapsed because “these early machines

suffered from high purchase prices, high maintenance costs, excessive noise, poor longevity, and

limited installation options.” By 1995, only two manufacturers remained, and sales sat at

approximately 2,000 units per year (Bodzin 1997). High first costs have been cited as a key

reason for the collapse of the market (Ashdown et al. 2004), and declining electricity prices (EIA

1996, EIA 2012e) certainly played a role in the diminishing economic viability of these HPWHs.

The Third Generation: 1999–2010

Facing a collapse of the HPWH market, Arthur D. Little, Inc., ECR International, and Oak Ridge

National Laboratory collaborated to develop a new integrated HPWH in the late 1990s to address

the aforementioned issues with the previous generation of products. At that time, the only

HPWHs on the market were add-on units to existing ERWHs. The first prototype was developed

All references to electricity price changes are relative to inflation (i.e., electricity price fluctuations are stated in

real terms).

in 1999, and within a year the final prototype was produced that had an EF of 2.47. The final

product, which was branded as the WatterSaver, was tested for durability and field performance

in subsequent years (Ashdown et al. 2004, Baxter and Linkous 2002, Tomlinson 2002).

During field testing of units across the United States, the average COP was 2, which is more than

double the efficiency of a comparable ERWH (Ashdown et al. 2004). In-house durability testing

at Oak Ridge National Laboratory uncovered no major durability issues (Baxter and Linkous

2002). Outside studies, however, had more mixed results from these products. A study in

California found an average COP of 1.27 for sites in California (Zogg, Murphy, and Hoyt 2004).

SWA performed field-testing of 20 WatterSaver HPWHs in 2002 in the northeast for

Connecticut Light & Power, which is a division of Northeast Utilities. With an average COP of

1.67, the efficiency of the unit was slightly lower for these field sites, and operational and

reliability issues with the units were uncovered (SWA 2004).

During the course of SWA’s monitoring, customer satisfaction was fairly high, and many

participants noted the dehumidification benefits. However, the study also identified some

consistent drawbacks with the daily operation of the systems. Many customers complained about

excessively hot water, and monitoring showed that water temperatures near the tops of the tanks

often exceeded 150°F. These high temperatures were partly due to excessive tank stratification—

water temperatures near the top could be 50°F higher than temperatures near the bottom—and in

many systems the high-temperature safety switches shut down the water heaters completely

(these were designed to shut down the system when temperatures reached 170°F). These issues

were communicated to the manufacturer (SWA 2004). Ultimately, the WatterSaver was removed

from the market because of the identified problems with installed performance and a nonexistent

service infrastructure (Environmental Building News 2005).

Current Products: 2010–Present

A variety of economic and regulatory factors are pushing HPWHs back into the mainstream

marketplace, as discussed in Section 1. Five major HPWH products are currently available in the

water heater market. Key specifications for these HPWHs are shown in Table 17. These

specifications include the two U.S. Department of Energy performance metrics: EF and FHR.

The EF represents the efficiency of the electric heating elements and tank losses under a

consistent, 24-hour test procedure. The FHR represents the amount of hot water that can be

supplied by a fully heated storage water heater during an hour of operation. Storage tank volume

is often used as a proxy for water-heating capacity, but FHR is the preferred metric. See Section

4.2 for more details about the EF and FHR test procedures.

RE is also used to quantify the efficiency of natural gas, fuel oil, and propane water heaters. RE is an approximate

metric of the burn efficiency of the heating device. Because electric resistance elements have an efficiency of 1, RE

is not tested or reported.

Table 17. Key Specifications of Integrated HPWHs Currently Available in the U.S. Market

Model

Capacity

(gal)

EF FHR (gal)

Electric Resistance

Elements

(kW)

GE GeoSpring

50 2.35 63

Upper: 4.5

Lower: 4.5

A.O. Smith Voltex

60/80 2.33 68/84

Upper: 4.5

Lower: 2

Stiebel Eltron Accelera 300

80 2.51 78.6

Upper: 1.7

Lower: None

Rheem EcoSense

40/50 2 56/67

Upper: 2

Lower: 2

AirGenerate AirTap

Integrated

50/66 2.39/2.4 60/75 Upper: 4

Outlook for the Future

While HPWH development in the United States focused on traditional refrigerants, Japanese

manufacturers, in conjunction with the Tokyo Electric Power Company, focused on developing

an HPWH based on the carbon dioxide (CO

) cycle. At the same time the WatterSaver was

being developed, a Japanese consortium started developing the CO

HPWH, which had a COP

of 3.4 when subjected to an ambient air temperature of 46.4°F and average tank temperature of

149°F. Marketed under the EcoCute brand, this CO

HPWH was released in 2001. The EcoCute

is a split system, meaning that it does not affect space-conditioning loads. Furthermore, the units

can be connected to a smart grid to try to move reheating periods to off-peak electricity periods.

COPs have increased to higher than 4 in more recent models, and sales of the EcoCute have

increased dramatically since its introduction (Hashimoto 2006).

The development of CO

HPWHs in the United States is just beginning. Market, technological,

regulatory, and structural hurdles still remain for this technology, but CO

HPWHs have been

actively pushed in recent years. Oak Ridge National Laboratory and GE have collaborated to

develop a prototype integrated HPWH based on the CO

cycle for the residential U.S. market.

Although this prototype is still in its early phases of development, future products may take

advantage of CO

as a more environmentally benign refrigerant (Abdelaziz et al. 2012).

Any omission of a manufacturer or product is unintentional, and no endorsement of any commercial product or

manufacturer is implied.

Appendix B: Alternative Water-Heating Calculations

Alternative water-heating products were evaluated using the equations underlying the

WATSMPL simulation program (EPRI 2000, 1995). For electric tanks the assumed recovery

efficiency is 1, and the pilot load is 0. The net water-heating load is computed using the hot

water load, inlet temperature, and outlet temperature:

( )

inoutpnet

TTVcQ −=

(9)

where

net

= net water-heating load (Btu)

V = hot water use (gal)

= specific heat of water (Btu/lb°F)

γ = specific weight of water (lb/gal)

out

= outlet temperature (°F)

= inlet temperature (°F).

The energy lost from the tank during the EF test Q

tank,EF

(Btu) is













−×= 1

EFtank,

water

(10)

where

water

= thermal energy drawn during EF test (41,063 Btu/day)

RE = recovery efficiency

EF = energy factor.

The energy lost from the tank during operation minus the lost energy from the pilot load Q

tank

(Btu) is

pilot

ambout

QRE

QQ ×−













−

×=

5.67135

EFtank,tank

(11)

where

pilot

= pilot load (Btu)

amb

= ambient temperature (°F).

The tank combustion losses Q

combustionloss

(Btu) is

( )













−×+= 1

tankloss combustion

QQQ

net

(12)

The gross water-heating load Q

net

(Btu) is

pilotnetgross

QQQ +++=

loss combustiontank

(13)

Appendix C: Monitored and Modeled Energy Use and Costs by

Site

Using the equations listed in Appendix B and the operating characteristics discussed in Section

5, the following site energy, source energy, and annual operating costs are listed in Table 18

through Table 21. HPWH numbers are annualized based on monitored data, and all other

systems are modeled using the operating conditions measured for each HPWH site.

Table 18. Annual Site Energy Consumption by Site

Site

HPWH

(kWh)

ERWH

(kWh)

Standard

Gas

(therms)

Premium

Gas

(therms)

Condensing

Gas

(therms)

Standard

Oil

(gal)

Premium

Oil

(gal)

Standard

Propane

(gal)

Premium

Propane

(gal)

1,155

3,045

173

149

117

124

106

189

163

1,969

5,899

304

271

221

217

193

331

295

1,838

4,152

222

195

157

158

139

242

213

1,186

2,981

169

145

115

120

104

184

159

3,612

4,739

258

225

180

184

161

281

246

892

2,502

149

126

106

162

137

714

1,756

116

126

103

909

2,403

144

122

103

157

133

1,169

2,665

154

132

103

110

168

144

1,475

3,434

198

169

133

141

121

216

185

3,080

5,927

303

271

222

216

193

330

296

1,380

3,114

176

151

120

125

108

192

165

1,317

2,393

145

122

104

158

133

922

2,235

136

114

148

125

Table 19. Annual Source Energy (MMBtu) Consumption by Site

Site

HPWH ERWH

Standard

Gas

Premium

Gas

Condensing

Gas

Standard

Oil

Premium

Oil

Standard

Propane

Premium

Propane

13.3

35.0

18.9

16.3

13.5

20.1

17.2

19.9

17.1

22.6

67.7

33.1

29.6

25.4

35.2

31.3

34.9

31.2

21.1

47.7

24.2

21.3

18.1

25.7

22.6

25.5

22.4

13.6

34.2

18.4

15.9

13.2

19.5

16.8

19.4

16.7

41.5

54.4

28.1

24.6

20.8

29.8

26.1

29.6

25.9

10.2

28.7

16.3

13.8

11.3

17.2

14.6

17.1

14.5

8.2

20.2

12.6

10.3

8.2

13.4

11.0

13.3

10.9

10.4

27.6

15.7

13.3

10.8

16.7

14.1

16.6

14.0

13.4

30.6

16.8

14.4

11.9

17.8

15.3

17.7

15.2

16.9

39.4

21.6

18.5

15.3

22.9

19.6

22.8

19.5

35.4

68.1

33.1

29.6

25.5

35.1

31.3

34.8

31.2

15.8

35.8

19.2

16.5

13.8

20.3

17.5

20.2

17.4

15.1

27.5

15.8

13.3

10.8

16.8

14.1

16.7

14.0

10.6

25.7

14.8

12.5

10.1

15.7

13.2

15.6

13.1

Table 20. Annual Operating Costs ($) by Site

Site

HPWH ERWH

Standard

Gas

Premium

Gas

Condensing

Gas

Standard

Oil

Premium

Oil

Standard

Propane

Premium

Propane

135

357

191

164

129

486

418

539

463

231

691

334

298

243

852

760

944

842

215

487

244

214

173

622

547

690

607

139

349

185

160

126

473

408

525

452

423

555

283

248

198

723

632

801

700

105

293

164

139

108

418

353

463

392

206

127

104

325

266

360

295

107

282

158

134

104

404

341

448

378

137

312

170

145

114

433

370

480

410

173

403

218

186

146

556

476

616

527

361

695

333

298

244

850

760

942

842

162

365

193

167

132

493

425

546

471

154

280

159

134

103

407

342

451

379

108

262

150

126

382

320

423

355

Table 21. Annualized Energy-Related Costs ($) by Site

Site

HPWH ERWH

Standard

Gas

Premium

Gas

Condensing

Gas

Standard

Oil

Premium

Oil

Standard

Propane

Premium

Propane

471

417

262

253

437

577

560

622

561

567

751

405

387

551

943

902

1,028

940

429

547

315

304

481

713

690

773

704

353

410

257

249

434

564

550

608

550

637

616

355

337

506

814

774

885

798

318

353

235

228

415

508

496

547

490

298

266

199

194

386

416

408

444

392

320

342

230

223

411

495

484

532

476

351

372

241

234

421

523

512

563

508

509

463

289

276

454

647

618

700

625

575

755

404

387

551

940

902

1,025

940

376

425

264

256

439

584

568

630

569

368

341

231

224

411

498

485

535

477

322

221

215

404

472

463

507

453

Appendix D: Identifying Heating Events

Choosing an appropriate analysis period to investigate the trends in the collected data is vital for

untangling the various input variables that can affect performance. Looking at the data on a

yearly site-by-site basis is inappropriate, because the sample size for this project is not large

enough and seasonal swings in operating conditions cannot be considered. Conversely, looking

at the 15-minute data is also inappropriate, because knowing the internal state of the tank, which

was not monitored in this study, would be necessary to understand the transient response of the

water heater to the operating conditions. Using daily data would be more appropriate, because

the system can be assumed to be in a steady state—i.e., the system operating variables do not

change considerably, but a larger sample size needs to be investigated.

Daily data, however, are not quite perfect, because the response rate of the heat pump system is

considerably slower than other heating mechanisms. A large draw at the end of one day can be

recovered entirely during the following day (Figure 22); the red ellipse highlights an incidence of

such a draw. As a result of this issue, the COP trends in relation to the draw volume are reversed.

One would expect the first day to have a lower COP than the second day because the draw

volume is smaller, which means that the standby losses account for a larger percentage of the

recovery load. The trend from using the daily data, however, is reversed, which confuses the

analysis.

Figure 22. Results from using daily data for analysis

120

160

200

240

8/20 8/21 8/22

Energy Used (Wh)

HW Used (Gallons)

HW Use

Heat Pump

Electric Resistance

20 gals

1.5 COP

35 gals

1.4 COP

A solution to this problem is to find periods of time when the water-heating tank starts fully

heated and ends fully heated. The criteria for finding such periods, which are defined as heating

events in this study, are: (1) the heating event must be at least 4 hours long, and (2) the heating

event must begin and end after a reheat cycle that has a duration one of at least 1 hour during

which no draws occurred. There are some tolerances to allow for noise in the measurements.

To demonstrate how this works in practice, consider Figure 23. Five heating events are displayed

in this graph, four of which have no electric resistance use. Each heating event begins and ends

at the end of a reheat cycle. As shown in the figure above, the COPs are related to the draw

volumes, as expected. Furthermore, the heating events have the advantage of operating near

steady state, the same as the daily values but without the limitations of the daily values.

Figure 23. Example of five heating events as defined in this study

Fri 02:30 Fri 12:45

Fri 20:00 Sat 02:45 Sat 15:45 Sun 17:15

HW Used (Gallons)

Fri 02:30 Fri 12:45

Fri 20:00 Sat 02:45 Sat 15:45 Sun 17:15

500

1000

1500

Energy Used (Wh)

HW Use Heat Pump Electric Resistance

60 gals

2.6 COP

27 gals

2.2 COP

53 gals

2.4 COP

44 gals

2.4 COP

70 gals

1.1 COP

Appendix E: Trifold Brochure for Consumers

DOE/GO-102016-4759 ▪ February 2016

buildingamerica.gov