A Solar-Assisted Heat Pump Water Heater

December 10, 2018
Winter 2018
A version of this article appears in the Winter 2018 issue of Home Energy Magazine.
Click here to read more articles about Hot Water

Water heating in residential buildings has become more important as the overall efficiency of buildings has improved over the last decade, with energy efficiency code requirements for A/C, lighting, and the building envelope. Modest improvement in energy efficiency for electric-storage water heaters was implemented in 2015, with the Code of Federal Regulations (CFR). But much greater improvements remain to be made.

Residential heat pump water heaters (HPWHs) save more energy than standard electric-resistance water heaters, because they are more efficient. HPWHs ranging from 50 gallons (189 liters) to 80 gallons (303 liters) are currently rated with a uniform energy factor (UEF) as high as 3.4, as indicated under the advanced-product list published by the Northwest Energy Efficiency Alliance (NEEA). And over the last ten years, PV conversion efficiencies have improved as costs have decreased. A design that integrated PV and HPWH would improve overall efficiency at low costs.

Tim Morrigan poses in front of two PV modules used for the system.

Carlos Colon makes microinverter connections.

As part of the effort to create zero energy buildings and high-efficiency water-heating systems, the Florida Solar Energy Center (FSEC)—under contract with the National Renewable Energy Laboratory (NREL)—has developed a PV-assisted heat pump water heater (PV-HPWH) prototype. This prototype combines two 310 watt-peak solar PV modules and grid-tied microinverters with a commercially available 189-liter HPWH. Using the HPWH’s integrated tank to store solar energy, the PV can be used with little grid interaction, thus avoiding net metering issues.

We began an evaluation of the PV-HPWH prototype in February 2016 at the FSEC hot-water systems laboratory in Cocoa, Florida. For testing, we utilized an automated hot-water load schedule totaling 59 gallons (223 liters) per day, which represents an average three-to-four-person family’s water use. We obtained an average coefficient of performance (COP) of 3.5 for the first month of February. COP is defined as the delivered hot water energy divided by the grid electrical energy used to produce the hot water. Larger COPs indicate higher efficiency. Further analysis excluding the PV contribution indicated a HPWH-alone performance COP of 2.1 in February, which is nearly identical to manufacturer claims for this first-generation unit.

The PV-HPWH prototype evolved with a series of implemented control improvements. Beginning in March 2016, we upgraded the system to automatically change the thermostat setting from 126°F (52°C) (baseline) to 140°F (60°C) depending on the power produced by the PV and the microinverters. The thermostat setting change was triggered by a minimum PV power threshold of 260 watts averaged over one minute. Following that change, during the month of April, the factory-supplied 4,500W bottom heating element was removed and replaced with a 750W element. The bottom heating element was then rewired and independently activated via an electronic controller and software developed at FSEC. The upper heating element was unchanged.

System Description

A diagram of the PV-HPWH as tested is shown in Figure 1. The HPWH used in the study is rated at 600 watts; however, depending on tank temperatures, the data indicated that the power draw ranged from 490 watts to 700 watts. A higher-than-rated power draw was routinely demonstrated as the compressor operated and heated water approaching the highest thermostat setting of 140°F (60°C) The fix-mount PV modules facing south at a 24° tilt produce the highest power during the midday hours (from 11 am to 2 pm). On average, the power produced by the PV modules was 369 watts during the midday period. The balance, or net energy, to operate the HPWH compressor came from the grid.

Figure 1. A diagram of the PV-heat pump water heater.

To match the energy produced by the PV modules to the electric-resistance-heating elements, we implemented power control circuitry using capacitive reactance, resulting in a two-stage electric resistance heating element. Further refinement of the developed controller led to techniques to maximize compressor operation efficiency and store most PV energy as hot water, with limited interaction with the grid and little waste of solar electricity produced. Once a full tank of 140°F (60°C) temperature water was reached and the compressor was shut off by the HPWH thermostat, additional heat energy was transferred into storage via the two-stage electric-resistance element. We tempered the hot-water delivery temperatures by using a mixing valve set at 126°F (52°C), which enabled the HPWH tank to act as thermal storage for the PV-generated electricity.

Increasing temperatures in the 189-liter tank from the 126°F (52°C) baseline to beyond 140°F (60°C) is the thermal storage equivalent of 2.2 kWh. Data indicated that thermal storage levels beyond 140°F (60°C) appeared during 75% of the days analyzed. Peak hot-water temperatures reached beyond 149°F (65°C) in May and August, as measured at the hot-water outlet port during the 3:49 pm hot-water draw, and averaged 144°F (62°C) in late afternoon. This allows the system to avoid any water-heating electrical demand during early evening hours—reducing the “duck curve” often associated with rapidly rising electrical demand from homes with PV systems during those hours.

On a daily basis, tank water heating in response to early morning cumulative hot-water draws (22 gallons or 83 liters) began around 7:30 am, but the heating process was interrupted at 8:30 am by a thermostat set up to 115°F (46°C). Hot-water recovery using the HPWH compressor was completed at 10:30 am, when solar resources are typically higher, by resuming a 126 gallon (52°C) thermostat set point. This process is illustrated in Figure 2, showing data and control thermostat set points recorded on August 23, 2016. During the example day, the compressor turns on at 7:38 am for only ten minutes, due to previous day storage and the 120°F (49°C) thermostat setting, and is followed by 192 watts of electric-resistance heating. The compressor resumes heating at 9:07 am as solar output increases and completes recovery by 11:30 am. The operation of the two-stage electric-resistance heating operation is visible in the afternoon, indicating extra energy being stored at a rate of 396 watts and 192 watts, respectively.

Figure 2. Hot-water recovery using the heat pump water heater compressor was completed at 10:30 am, when solar resources are typically better.

Figure 3. Test results from the prototype PV-heat pump water heater operating in Central Florida.


Figure 3 illustrates test results from the prototype PV-HPWH operating in central Florida. The bar graph shows the daily COP of the overall system for the month of September 2016, as defined by the following relationship:

COP = Hot water energy out

Grid electrical energy in

Figure 4 shows the daily solar radiation from the two PV modules (dotted line) in watt-hours per square meter per day on the right axis. COPs around 6 are typically achieved on days with daily plane-of-array solar radiation above 4 kWh per square meter. Variation in the COP on any particular day is affected not only by the solar irradiance, but also by its distribution against loads on that particular day. Efficiency is further influenced by the solar irradiance on the preceding day, as this determines the stored thermal energy over evening hours. Overnight tank thermal storage is particularly important for the efficiency of serving early morning hot-water draws.

Figure 4. Measured long-term performance recorded through January 2017.

Figure 4 also shows measured long-term performance recorded through January 2017. Average monthly COP is shown on the left y-axis, and kWh per day of electricity consumption is shown on the right y-axis. Performance of the PV-HPWH has been exceptional, demonstrating average monthly COPs as high as 6.6 and 7 for the months of May and July respectively. COPs leveled off at around 6 for the months of August through October and declined in November. Further improvement could be realized by utilizing a portion of the 18% of total energy produced by the PV/microinverters that is unused and is fed back into the grid during the early morning and late afternoon hours. Future development work aims to maximize PV-supplied energy to be used by the system, and to island the electric power if desired so that there is no interaction with the grid.

Performance data for the months of December and January 2017 reveal COPs of 5.1 and 4.8 respectively. These COPs are consistently higher than the COPs for February 2016, when controls optimization and the two-stage heating element were not in place.

We also performed a time-of-day analysis to determine the hourly peak-demand reduction potential of the PV-HPWH against standard electric-resistance water heating. Figure 5 presents the hourly demand as compared to a 50-gallon (189-liter) standard electric water heater (red line) that was operated simultaneously in the laboratory.

Figure 5. We performed a time-of-day analysis to determine the hourly peak-demand reduction potential of the PV-heat pump water heater against standard electric-resistance water heating.

The plot also shows the diversified water-heating-only demand profile (black dashed line) of 60 residential electric water heaters operating in Florida homes. These water heaters were recently monitored as part of the DOE Building America Phased Deep Retrofit (PDR) study. Because it stores an extra ~2.2 kWh of thermal energy, the PV-HPWH would not increase the late afternoon ramp-up demand on utility generation caused by the PV systems’ decreasing electricity production.

Table 1 summarizes the PV-HPWH’s performance for the 12 months from February 2016 through January 2017. Analysis performed on the data after May 2016, when the auxiliary heating by electric resistance was implemented for additional heat storage, indicated that the solar contribution from the PV and microinverters averaged 65.6% of the total electricity used by the system. The average long-term COP of 5.4 was exceptional.

Table 1. Summary of PV-HPWH performance (February 2016–January 2017)

Average Monthly Daily Electric Consumption Average Monthly COP Average PV Energy Generated Added Storage Above 126°F (52°C) Average Hot Water Max Temp Stored Average Daily Hot Water Delivered (w/126°F Mix Valve Setting)
1.2 kWh/day 0.7–2.1 (max-min) kWh/day 5.4 kWh/ day 4.5/7.0 (max–min) kWh/day 2.3 kWh/day 2.1 kWh/day 2.1 kWh/day 57 gallons
(215 (liters)
211,868 kJ

The unit modified to a PV-HPWH was a first-generation HPWH with a UEF of 2.45. Many second-generation HPWHs have UEFs of 3.1–3.7. It should be noted that the COPs described here are actually much more impressive than they seem at first, since the UEF test procedure generally allows tested HPWHs to avoid use of the electric-resistance heating elements, which can significantly degrade real-world performance.

Performance data collected from the PV-HPWH to date show that this is the highest-efficiency electric water-heating system ever tested under FSEC’s hot-water system evaluation program, which was conducted under the DOE Building America program through 2016. Figure 6 places the PV-HPWH at the top of the chart, with an average grid electricity use of 1.2 kWh per day. Note that the conventional 50-gallon (189-liter) electric-resistance tank used about 7.6 kWh per day with an integrated annual COP of approximately 0.8. Accordingly, the prototype PV-HPWH saved 83% of the energy typically needed for conventional electric-resistance water heaters.

Figure 6. The prototype PV-heat pump water heater saved 83% of the energy typically needed for conventional electric-resistance water heating.


The prototype PV-HPWH showcases innovative strategies for distributed PV systems that limit the grid interaction with water heating, and provides increased thermal energy storage. Sophisticated controls have been modified through experimentation such that higher tank temperatures are achieved during the day when solar availability is high without triggering the compressor during early morning hot-water draws. The system utilizes a custom appliance control module interface to vary thermostat settings (115°F to 140°F, or 46°C to 60°C) depending on time of day and solar radiation levels. It also prioritizes thermostat settings on a time-of-day basis. By altering the thermostat down to 115°F (46°C) during early morning draws, the system can disrupt compressor heating recovery normally set to 126°F (52°C) and shift the remainder of recovery to times when adequate solar resources are typically available (after 10:30 am). When tank temperatures are satisfied, the remaining PV electricity is stored in the tank using staged electric-resistance elements. Typical performance sees hot-water storage greater than 149°F (65°C) at sunset, but is limited at that point. A mixing valve provides hot water at the target temperature (126°F, or 52°C). By dynamically altering tank temperature, an equivalent of ~2 kWh of electrical energy is stored for use during evening hours. Typically, there is no water-heating grid electricity demand during utility summer peak periods. Thus this system reduces the “duck curve” that results from many residential PV installations.

The prototype PV-HPWH system equipment had a retail cost of $2,053, including the HPWH, PV, microinverters, controls, and tempering valves. This is significantly less than traditional solar-thermal systems, which average $6,000–8,000. Key advantages of the PV-HPWH technology include

  • simplified installation, which is likely to reduce installed costs;
  • no need for plumbing or pumps; no need for freeze protection;
  • PV output at given irradiance higher under cold conditions, when water-heating loads are higher;
  • solid-state components, leading to greater long-term reliability; and
  • a PV-to-thermal storage strategy that typically produces 2.2 kWh of evening load shift relative to standard electric-resistance systems.

Analysis of data from the original prototype suggests even better performance is possible. We anticipate further refinement of the controls within a heuristic control framework (learning prevailing household hot-water load and weather patterns). Also, the latest generation of HPWH compressors operate at a power level of ~50 watts less than the unit used in our demonstration. Finally, in northern climates, 303-liter hot-water storage tanks could be utilized along with additional PV modules to further improve performance.

learn more

See the full report upon which this article is based, including modeling of the PV-HPWH across various climate zones.

Download the advanced-product list published by the NEEA.

For information about the 2013 DOE Building America PDR study, see Parker, Danny, et al. Retrofits in Existing Homes in Florida Phase I: Shallow and Deep Retrofits, FSEC-CR-2018-16. Final Report prepared by Florida Solar Energy Center, February 2016.

FSEC’s hot-water system evaluation program, conducted under the DOE Building America program through 2016, found that the newer General Electric HPWH, the GEH50DFEJSRA (more recently produced by Bradford White), has a 3.25 UEF with a compressor power of 550 watts.

To evaluate the potential of PV-HPWH technology across North America, we performed a simulation analysis using a calibrated TRNSYS model of the system. This simulation very accurately described the laboratory performance seen in Florida. With the PV-HPWH located inside the conditioned space, performance was estimated across 19 highly diverse North American climates. The results showed strongly improved performance over standard HPWH systems. A two-module PV-assisted system appeared adequate in warmer climates, but a three-module system was best in colder climates. System COPs ranged from 2.4 to 3.7 across climates in the two-module system. The use of newer lower-wattage compressors would probably improve upon these simulation results.

We made a preliminary evaluation of relative economics, using an installed-system cost estimate along with state-level electricity prices. Including provision for the federal solar tax credit on the solar element, the estimated simple paybacks were 6–11 years. Economics were best in locations with the highest electricity costs, such as Hawaii, California, and New York. However, economics were positive in all locations. Finally, an evaluation of relative greenhouse gas emissions showed that the PV-HPWH technology could significantly reduce emissions over electric and natural-gas water heating of all types across all 19 regions. A pilot demonstration of the PV-HPWH technology in a residential water-heating field project seems the next logical step.

Refinement based on results could lead to development of a new generation of residential water-heating equipment addressing the combined needs for very high-efficiency and effective energy storage while greatly reducing associated greenhouse gas emissions.

Danny Parker and Carlos Colon are research scientists at the Florida Solar Energy Center. Tim Merrigan and Jeff Maguire perform research on buildings at NREL.

  • 1
  • NEXT
  • LAST
Click here to view this article on a single page.
© Home Energy Magazine 2020, all rights reserved. For permission to reprint, please send an e-mail to contact@homeenergy.org.
Home Energy Pros
Discuss this article with other home performance professionals at Home Energy Pros!.

Add a new article comment!

Enter your comments in the box below:

(Please note that all comments are subject to review prior to posting.)


While we will do our best to monitor all comments and blog posts for accuracy and relevancy, Home Energy is not responsible for content posted by our readers or third parties. Home Energy reserves the right to edit or remove comments or blog posts that do not meet our community guidelines.

Related Articles
SPONSORED CONTENT Insulated, Air-Sealed Drapes Learn more! Watch Video