| Table 1: Basic Heat Pump and Home Information | |||||||
|---|---|---|---|---|---|---|---|
| Heat Pump ID | Home Location | Heat Pump Info | Avoided Fuel Type | Home Information | |||
| Model | HSPF Rating | Floor Area, ft2 | Shell Efficiency | Building Type | |||
| 1 | Seward | Fujitsu 24K (2-heads) | 10.5 | #1 Oil | 2,077 | Mid | 2-story |
| 2 | Seward | GREE Sapphire 12K | 14.0 | #1 Oil | 1,268 | Low | 1-story on crawl |
| 3 | Seward | Fujitsu 18RLXFZ | 9.3 | Spruce Wood | 1,536 | Low | 1-story 1/2 basement |
| 4 | Seward | GREE Sapphire 9K | 15.0 | Spruce Wood | 764 | Mid | 1-story on crawl |
| 5 | Seward | Fujitsu 12LMAS1 | 12.5 | #1 Oil | 1,400 | Mid | 1-story, insulated floor |
| 6 | Seward | Chiltrix CX34, air-to-water | 13.0 | Electricity | 2,753 | Super | 2-Story w/garage |
| 7 | Seward | Fujitsu AOUH18LPAS1 | 11.0 | Spruce Wood | 1,668 | High | 1.5-Story |
| 8 | Seward | GREE Shappire 12K | 14.0 | #1 Oil | 2,407 | Low | 1.5 Story w/ basement, log |
| 9 | Seward | Fujitsu 12K | 13.8 | Spruce Wood | 1,872 | Mid | Split/Tri-level |
| 10 | Seward | Fujitsu AOU45 RXLFZ | 10.3 | Propane | 1,400 | Mid | 2-story, New |
| 11 | Homer | Suhoer Shaw 12 kW, air-to-water | 12.0 | Spruce Wood | 2,215 | High | 1-story |
| 12 | Seward | GREE Sapphire 9K | 15.0 | Electricity | 480 | Super | 2-story |
| 13 | Seward | Suoher Shaw 6 kW, air-to-water | 14.0 | Electricity | 832 | Super | 1-story |
| 14 | Seldovia | Fujitsu AOU24RLXFWH | 10.5 | Spruce Wood | 850 | High | 2-story Yurt |
| "HSPF Rating" is a standard heating efficiency rating for the heat pump. | |||||||
| For heat pump models identified as "air-to-water", HSPF rating values are not available but are estimated here. | |||||||
| "Shell Efficiency" is our qualitative assessment of the insulating level of the building shell. | |||||||
Kenai Peninsula Heat Pump Monitoring Project
Date: May 10, 2024
Authors: Alan Mitchell, Analysis North (alan@analysisnorth.com) and Phil Kaluza, Arctic Energy Systems (pkaluza@gmail.com)
This work was funded by the Alaska Housing Finance Corporation
Executive Summary
This study monitored the electrical power use of 14 residential heat pumps located on the Kenai Peninsula of Alaska. All but two of the units were located in the Seward, Alaska area; the remaining two were in Homer and Seldovia. Electric power was recorded in 10-minute intervals, and more than one year of data was collected for most of the heat pumps. The heat pumps were primarily ductless (mini-split) heat pumps, although three of the units were air-to-water heat pumps that served both the space and domestic hot water end uses.
For each heat pump, electric power use was correlated with outdoor temperature. By applying this relationship to typical annual weather data for each site, an estimate of typical annual electricity use was made. Typical annual electricity use across the units varied from 1,730 kWh/year to 20,290 kWh/year (large air-to-water heat pump), with a median value of 4,200 kWh/year. The peak electrical demand from the heat pumps (2-hour average values) varied from 1.1 kW to 4.2 kW, with a median value of 2.0 kW.
Relying on the algorithms used in the Alaska Mini-Split Heat Pump Calculator estimates were made for the fuel savings that resulted from the use of each heat pump. It is important to understand that this study did not measure heat output or fuel savings for each heat pump; instead, those values were estimated from the HSPF efficiency rating of the heat pump, and hourly outside temperature values. The fuel types avoided by use of the heat pumps included #1 heating oil, wood, electricity, and propane. Actual fuel saving values are provided in Table 3 of the main report.
Annual net cost savings (considering the cost of the electricity required to run the heat pump) varied from -$831/year to $2,689/year, with a median value of $423/year. The negative value resulted from avoiding wood heat produced by an efficient wood boiler in the Homer Electric Association service territory, where electric rates are relatively high. No cost was attributed to the homeowner labor required to use wood heat. The highest heat pump cost savings occurred for homes where the alternative source of heat was electricity or propane. Cost savings are summarized in Table 3 of the main report.
Introduction
This study monitored the electrical power use of 14 residential heat pumps located on the Kenai Peninsula of Alaska. More than one year of data was collected for most of the heat pumps. In addition to electrical power use data, outdoor temperature data from nearby weather stations was matched to the power consumption data. A number of conclusions were derived from the data, including estimates of annual electricity use, amounts of avoided heating fuel, electrical peak demand levels, and qualitative conclusions regarding usage patterns of the heat pumps. These results and conclusions are summarized in the rest of the report.
Study participants were recruited by the authors; many were personal friends, and the local Seward heat pump installer also identified willing participants. The sample was clearly not a pure random sample of installations, but it did span a wide variety of conditions. The homes ranged in size from 480 - 2,753 square feet. Many heat pumps only served a portion of the heat load in the home, but a few of the heat pumps served the entire space heating load. Various backup heating sources were present in the sample, including wood, oil (Toyostoves, boilers, and furnaces), and electric resistance heat.
A custom-made electrical power sensor that transmitted it’s power readings via long-distance LoRaWAN wireless technology was used to measure the electrical use of most of the heat pumps. A number of LoRaWAN gateways to receive the wireless transmissions were already present in Seward where the bulk of the participants were located. Here is picture of one of the installations; the power monitor is pointed to by the red arrow:
Power was measured continuously, and an average was computed and transmitted every 10 minutes.
Home, Heat Pump, and Data Collection Information
Table 1 below gives some basic information about the homes and heat pumps that participated in the study.
Table 2 below gives information on the dates that electric power data was collected from each heat pump. “Months of Valid Data” only counts the days when valid power measurement data was collected.
| Table 2: Dates of Collected Data | |||
|---|---|---|---|
| Heat Pump ID | Months of Valid Data |
First Date | Last Date |
| 1 | 14.3 | Feb 18, 2023 | May 9, 2024 |
| 2 | 12.5 | Feb 18, 2023 | Apr 11, 2024 |
| 3 | 14.6 | Feb 18, 2023 | May 9, 2024 |
| 4 | 14.3 | Feb 18, 2023 | May 9, 2024 |
| 5 | 8.9 | Feb 18, 2023 | May 9, 2024 |
| 6 | 14.6 | Feb 18, 2023 | May 9, 2024 |
| 7 | 14.5 | Feb 18, 2023 | May 9, 2024 |
| 8 | 14.8 | Feb 15, 2023 | May 9, 2024 |
| 9 | 14.8 | Feb 15, 2023 | May 9, 2024 |
| 10 | 14.4 | Feb 18, 2023 | May 9, 2024 |
| 11 | 12.7 | Mar 12, 2023 | May 9, 2024 |
| 12 | 12.7 | Apr 20, 2023 | May 9, 2024 |
| 13 | 5.5 | Jun 1, 2023 | Apr 8, 2024 |
| 14 | 12.9 | Jan 6, 2023 | Apr 7, 2024 |
| Days with missing data are excluded from the "Months of Valid Data" quantity. | |||
Electricity Use and Avoided Fuel
Primary goals of the study were to estimate the annual electricity use of each heat pump and estimate how much heating fuel was saved by each heat pump during a typical year. The measurement period for each heat pump was not exactly one year, and the monitoring period may not have been representative of normal weather occurring during a typical year. To address these issues, the actual data collected was used to create a model of electric power use versus outdoor temperature. The data collected did span a wide range of outdoor temperatures, so these models give a reasonable estimate of the average power consumption of the heat pump across a broad range of outdoor conditions. These models are presented for each heat pump in the last section of this report.
To estimate annual electricity use of each heat pump during a typical year, the electric power versus outdoor temperature model was applied to temperature data for the three complete years of 2021 - 2023. Electric power use was estimated for each hour of those three years, and the yearly electricity consumption values were averaged. Those average annual consumption values are deemed typical and are shown in Table 3 below.
To estimate how much heating fuel was saved by use of the heat pump, the COP (efficiency) of the heat pump was estimated for each of the hours in the 3-year period. The algorithm used in the Alaska Mini-Split Heat Pump Calculator was used. The inputs to that COP estimation model are the HSPF rating of the heat pump and the outdoor temperature. With the COP value, the avoided heating load can be determined, and with an estimate of the efficiency of the heating system that would be used in the absence of the heat pump, an estimate of avoided fuel can be made. The Net Annual Cost Savings column in Table 3 below estimates the dollars per year saved by the heat pump, considering the savings of heating fuel offset by the cost of electricity to run the heat pump.
Finally, an annual average heat pump COP (efficiency) is estimated for each heat pump, accounting for the variation in efficiency and electric power use of the heat pump across the year.
It is important to emphasize that this study did not measure the output or efficiency of each heat pump. The two values that were measured were the input electric power to the heat pump and the outdoor temperature. Heat pump efficiency and avoided fuel were estimated given the HSPF efficiency rating of the heat pump, efficiency of the backup heating system, and outdoor conditions. Given the uncertainty in these estimations, there is noticeable uncertainty in the stated avoided fuel quantities and annual average COP. There is substantially less uncertainty in the heat pump annual electricity use, because the only estimation involved in this quantity was the adjustment from weather conditions during the monitoring period to weather conditions during a typical year.
Here are the results:
| Table 3: Estimated Annual Heat Pump Electricity Use and Fuel Savings | ||||
|---|---|---|---|---|
| Heat pumps are grouped by the type of fuel they avoid. The numeric labels for each row are the Heat Pump IDs. | ||||
| Annual Heatpump Electricity, kWh |
Avoided Fuel Quantity |
Annual Heat Pump COP |
Net Annual Cost Savings |
|
| #1 Oil | ||||
| 1 | 5,537 | 391 gallon | 2.3 | $459 |
| 2 | 6,065 | 501 gallon | 2.7 | $812 |
| 5 | 2,548 | 191 gallon | 2.5 | $249 |
| 8 | 7,536 | 614 gallon | 2.7 | $937 |
| Spruce Wood | ||||
| 3 | 4,726 | 3.82 cords | 2.1 | $201 |
| 4 | 2,293 | 2.37 cords | 2.7 | $264 |
| 7 | 3,673 | 3.16 cords | 2.3 | $219 |
| 9 | 2,706 | 3.03 cords | 3.0 | $387 |
| 11 | 13,667 | 8.98 cords | 2.4 | −$831 |
| 14 | 2,737 | 2.37 cords | 2.3 | $12 |
| Electricity | ||||
| 6 | 7,887 | 20,286 kWh | 2.6 | $2,689 |
| 12 | 1,731 | 4,609 kWh | 2.7 | $624 |
| 13 | 3,398 | 9,228 kWh | 2.7 | $1,264 |
| Propane | ||||
| 10 | 5,774 | 602 gallon | 2.2 | $1,649 |
| For heat pumps avoiding electric resistance heat, the net electricty savings can be calculated by subtracting the Annual Heat Pump Electricity from the Avoided Fuel Quantity. | ||||
The graph below shows the distribution of heat pump annual electricity use across all of the heat pumps:
Average: 5,020 kWh / year, Median: 4,199 kWh / year
The graph below shows the distribution of net annual cost savings across all of the heat pumps:
Average: $638 / year, Median: $423 / year
The largest estimated cost savings were realized by homes where electricity resistance or propane were the alternative heat source. Those fuels have high cost per delivered BTU of heat. Homes avoiding wood heat had low or negative dollar savings, as the electricity cost of using a heat pump is close to cost of heating with wood, particularly for heat pumps served by Homer Electric Association, which had higher electric rates than Seward Electric during the study period. However, many of the wood heat study participants were motivated to reduce the homeowner labor required to heat with wood and also wanted a form of heat that did not require human participation while household members were away from the home.
Significant carbon emission reductions were achieved by all heat pump installations, due to carbon emissions from the electrical generation supplying the heat pump being less than the carbon released by using alternative heating fuels. This statement is true even for the wood heat homes when the emissions analysis is confined to a 20-year period. However, no quantitative carbon calculations are presented in this report.
The detailed inputs involved in the avoided fuel and cost savings calculations are presented in Table 4 below. Note that fuel and electricity prices comes from the AkWarm fuel/utility price survey that was published March 1, 2024. Prices include sales tax. Spruce wood prices were not present in the survey but were estimated to be $300/cord + sales tax for split wood. Since the AkWarm price survey, electricity rates in Seward have been increased substantially, but that increase is not reflected in the study results.
| Table 4: Assumptions related to Fuel and Cost Saving Estimates | ||||
|---|---|---|---|---|
| Heat Pump ID | Avoided Fuel Type | Efficiency of Backup Heat System |
Avoided Fuel Price / Unit |
Electricity Price $/kWh |
| 1 | #1 Oil | 80% | $4.25 | $0.217 |
| 2 | #1 Oil | 80% | $4.25 | $0.217 |
| 3 | Spruce Wood | 50% | $321.00 | $0.217 |
| 4 | Spruce Wood | 50% | $321.00 | $0.217 |
| 5 | #1 Oil | 82% | $4.09 | $0.209 |
| 6 | Electricity | 100% | $0.22 | $0.217 |
| 7 | Spruce Wood | 50% | $321.00 | $0.217 |
| 8 | #1 Oil | 82% | $4.09 | $0.209 |
| 9 | Spruce Wood | 50% | $321.00 | $0.217 |
| 10 | Propane | 80% | $4.82 | $0.217 |
| 11 | Spruce Wood | 70% | $309.00 | $0.264 |
| 12 | Electricity | 100% | $0.22 | $0.217 |
| 13 | Electricity | 100% | $0.22 | $0.217 |
| 14 | Spruce Wood | 50% | $309.00 | $0.264 |
Heat Pump Peak Electrical Demand
The electric power data from each heat pump was also analyzed to determine the peak electric draw of the heat pump. As stated before, average electric power data was recorded in 10-minute intervals. For these 10-minute intervals, Table 4 shows the electric power demand for the highest interval. The table also shows the peak demand for a 2-hour averaging interval. This is perhaps a more useful measure for electric utilities, because individual, short-term heat pump power peaks are unlikely to coincide in time. Looking at 2-hour averaged intervals helps account for this diversification of peak demands.
| Table 5: Peak Electrical Demand from Heat Pump | |||||
|---|---|---|---|---|---|
| Shown for two different lengths of averaging intervals: 10 minutes and 2 hours. | |||||
| Heat Pump ID | 10-minute Peak Demand | 2-hour Peak Demand | |||
| Peak kW | When | Peak kW | When | ||
| 1 | 2.5 | 2023-11-19 15:24 with Outdoor Temperature = 19.3 °F | 2.3 | 2024-01-26 04:00 with Outdoor Temperature = 12.7 °F | |
| 2 | 2.1 | 2023-11-20 05:17 with Outdoor Temperature = 13.7 °F | 2.0 | 2023-03-05 06:00 with Outdoor Temperature = 19.9 °F | |
| 3 | 2.0 | 2024-02-03 00:55 with Outdoor Temperature = -2.0 °F | 2.0 | 2024-02-01 12:00 with Outdoor Temperature = 7.7 °F | |
| 4 | 1.6 | 2023-03-14 07:23 with Outdoor Temperature = 16.0 °F | 1.1 | 2023-12-08 08:00 with Outdoor Temperature = 7.3 °F | |
| 5 | 2.7 | 2023-12-28 19:48 with Outdoor Temperature = 17.6 °F | 1.5 | 2023-12-30 00:00 with Outdoor Temperature = 11.2 °F | |
| 6 | 3.1 | 2024-02-25 00:09 with Outdoor Temperature = 28.0 °F | 2.9 | 2023-11-13 22:00 with Outdoor Temperature = 30.0 °F | |
| 7 | 2.6 | 2024-01-31 06:47 with Outdoor Temperature = 3.9 °F | 2.5 | 2024-02-03 08:00 with Outdoor Temperature = -0.0 °F | |
| 8 | 2.2 | 2024-01-31 17:04 with Outdoor Temperature = 2.7 °F | 2.1 | 2024-01-31 14:00 with Outdoor Temperature = 6.8 °F | |
| 9 | 2.7 | 2023-03-19 12:56 with Outdoor Temperature = 32.8 °F | 2.0 | 2023-11-06 00:00 with Outdoor Temperature = 30.9 °F | |
| 10 | 4.6 | 2023-03-03 20:06 with Outdoor Temperature = 21.7 °F | 4.2 | 2023-03-03 20:00 with Outdoor Temperature = 21.7 °F | |
| 11 | 4.5 | 2023-03-21 15:07 with Outdoor Temperature = 31.9 °F | 4.0 | 2023-03-22 14:00 with Outdoor Temperature = 32.6 °F | |
| 12 | 1.9 | 2023-12-29 17:57 with Outdoor Temperature = 14.0 °F | 1.4 | 2023-12-22 06:00 with Outdoor Temperature = 11.4 °F | |
| 13 | 2.3 | 2023-08-16 11:59 with Outdoor Temperature = 52.3 °F | 1.7 | 2023-06-02 20:00 with Outdoor Temperature = 46.0 °F | |
| 14 | 1.6 | 2024-03-01 04:15 with Outdoor Temperature = 16.1 °F | 1.4 | 2024-03-02 04:00 with Outdoor Temperature = 19.1 °F | |
| Air-to-water heat pumps often serve substantial thermal mass. Peak demands for those systems can occur during warmer outdoor conditions due to increases in the temperature setpoint of the system. This is the explanation for the peaks shown for Heat Pump ID 13. | |||||
The plot below shows each heat pump’s peak demand, both for the 10-minute averaging interval and the 2-hour averaging interval.
| Table 6: Peak Demand Statistics | ||
|---|---|---|
| Interval Size | Average | Median |
| 10-Minute | 2.6 | 2.4 |
| 2-Hour | 2.2 | 2.0 |
Heat Pump Power use versus Outdoor Temperature
This section presents a graph for each heat pump showing the relationship of the heat pump electric power use to the outdoor temperature. Each dot on the graphs below contains one day of heat pump data. The horizontal position of the dot is the average outdoor temperature during that day, and the vertical position is the average electric power use in Watts. The line on the plot shows the trend in data. Average temperature and and power for each 6 °F outdoor temperature interval are used to create the line. For the creation of the line, there is no assumed mathematical relationship between heat pump power and outdoor temperature; the line is simply formed by averaging data points together in each 6 °F outdoor temperature bin.
If a heat pump were used to serve the entire space heating load of a building, the relationship between outdoor temperature and heat pump electrical power use would have the following shape:
The heating load for a home increases linearly as outdoor temperature declines. However, the curve above shows non-linear, upward curvature as temperature declines. This is due to the fact that the efficiency of the heat pump gets worse as outdoor temperature declines.
This general shape is present in many of the graphs below, indicating that heat pump is covering a large portion of the heating load of the space being served by the heat pump. However, for a number of the other heat pumps, the graphs do not show the upward curvature at cold temperatures. For these homes, a secondary heating system is being more heavily utilized at colder temperatures.
Some participants used their heat pumps to provide air-conditioning on warm days. Some of the power versus outdoor temperature graphs below reflect that increased power use at warm outdoor temperatures. For example, see heat pumps #9 and #10.
Note that the horizontal axis range on each graph below is the same, but the vertical axis range varies by heat pump. Heat pump size and usage varies substantially across these homes.
This heat pump was out of service from Jan 5 - Feb 6, 2024, a period with substantial cold weather. Thus, there is little data for outdoor temperatures below 10 °F. During cold weather, the oil forced-air furnace is used to supply sufficient heat to the back bedrooms.
Both a Toyostove and wood stove were used as alternative heat sources. Sometimes overlap in thermostat settings between the Toyostove and the heat pump occurred, causing both to operate simultaneously on Low.
A major power outage caused a heat pump circuit board failure and failure of the power monitoring system. Monitoring halted in mid-January 2024 due to this issue.
Heat pump provided nearly all the heat to this well-insulated home, with only occasional use of the wood stove.
The heat pump capacity in this old, leaky log home is less than the heat load for outdoor temperatures less than about 35 °F. This explains why the power curve noticeably flattens below that temperature. Also, conflicts in use between the Toyostove and the heat pump occurred.
The owner burns substantial wood during the colder periods and would then rely on the heat pump to cover the warmer days, as it was better at controlling the indoor temperature than the wood stove.
Substantial wood boiler heating use occurs in this home at temperatures below 20 °F.
Small, efficient two-story apartment with unit installed on the main floor. The heat pump serves the entire space heating load. Tenant would set back the temperature during the day while working.
This air-to-water heat pump serves the entire space heating load and also serves most of the domestic hot water load. Due to the domestic hot water load, you can see noticeable power use at warm temperatures.
The heat pump was turned off during a cold period during the winter of 2023-24 while the occupants were away on vacation. Thus, there is a lack of cold weather data.
