The analysis results presented here used the following approach and input assumptions. Differences in approach or format of results may make comparisons to other efficiency potential analyses invalid. More details on the methodology and input assumptions can be found in the NREL technical report, Energy Efficiency Potential in the U.S. Single-Family Housing Stock.

Definition of energy efficiency potential

The energy efficiency potential is presented as annual energy savings specific to each state, assuming full turnover of the stock of equipment and appliances at wear out, which could take 15–30 years, depending on the type of equipment. Full uptake of applicable and cost-effective enclosure improvements was also assumed. Thus, the savings presented are the economic potential, as opposed to market potential which might be informed by historical adoption rates. Policy and programs could be implemented to lower other market barriers and reduce the difference between market potential and economic potential.

Energy savings

The energy savings estimates were calculated using ResStock™, a highly granular model that uses 350,000 physics-based building energy models (OpenStudio^®/EnergyPlus™) to statistically represent the diversity of the U.S. single-family housing stock (80 million homes) across a range of climates (216 climate regions), vintages, sizes, fuel types, equipment, insulation, occupancy, etc. For heating/cooling equipment and appliance upgrades considered at wear out, only the incremental savings and cost over the reference replacement scenario (e.g., federal minimum standards) is counted. Detailed descriptions of each improvement scenario are provided in the technical report. The analysis did not account for state/local codes and standards.

Packages/measure interaction

The statewide savings potential values presented in the upper right box are based on simulated packages of cost-effective improvements, including the state's top 10 improvements listed and other, lower priority, improvements. The packages are tailored to maximize the NPV in each of the 350,000 representative home models. The packages account for any diminished returns due to interaction between heating/cooling equipment and enclosure upgrades, as well as the potential for reduced upfront cost through heating/cooling equipment downsizing. The green bar graph shows the statewide economic potential energy savings of the top 10 individual improvements, not accounting for interactions between them.

Cost-effectiveness perspective

The presented energy efficiency potential aggregates savings across all homes in which the improvement or package has a positive net present value (NPV), evaluated using costs and benefits from the building owner’s perspective rather than a utility or societal perspective. Health and safety benefits were not quantified.

Economic assumptions

For NPV calculations, 30 years of future cash flows (utility bill savings, equipment replacement at end of life, and residual value) are brought to the present using a 3% real discount rate. The technical report provides additional results using an alternative (more stringent) cost-effectiveness criterion of simple payback period less than five years.

Tenant-occupied homes

The same economic calculations are used for both owner-occupied and tenant-occupied single-family homes. For tenant-occupied housing, it is assumed that either the building owner pays the utility bills or rent is increased by an amount equal to utility bill savings.

Utility prices

Electricity rates were assumed to be flat volumetric charges ($/kWh), were derived from EIA Form 861 sales and revenue data (2013), assumed $8 per month in fixed customer charges, and were applied at the county level based on utility service territories reported on EIA Form 861. Rates for on-site fuel use were derived from 2015 EIA data (natural gas) and 2010 EIA data (propane and fuel oil), and were applied by state.

Improvement costs

Improvement costs include all material, labor, and overhead costs paid by the building owner. In the case of equipment replaced at wear out, only the incremental cost of the improvement over the reference (e.g., federal minimum standards) is included. As noted above, future replacement costs and residual value at the end of the 30-year cash flow analysis are accounted for. With a few exceptions, the initial costs and lifetimes for each upgrade and reference scenario are national averages sourced from the National Residential Efficiency Measures Database.

Incentives

State, utility, and local incentives (e.g., rebates) were not included in the economic analysis, due to the large number of unique incentives that exist. A federal income tax credit for residential energy efficiency was included and assumed to be available in future years (capped at $500 per household). The eligible improvements include high-efficiency heating, cooling, and water heating equipment, as well as insulation and windows upgrades.

Sectoral scope

The analysis covers single-family detached housing only. The housing stock characteristics tool developed for ResStock currently is limited to single-family detached housing and excludes all multifamily buildings (including duplexes and townhomes) as well as mobile homes. Expansion to the multifamily sector is planned in 2018.

Geographic scope

The fact sheets are limited to the 48 contiguous U.S. states. Sources of housing characteristics and consumption data (particularly the Residential Energy Consumption Survey) for Alaska, Hawaii, and U.S. territories tend to have low sample sizes, resulting in high uncertainty in the data.

Temporal scope

House counts and housing characteristics are a snapshot based on circa-2012 data; projections of future construction and changes in housing characteristics were not included in this analysis.

The ResStock software can be used to analyze additional scenarios with different input assumptions, for additional technologies, or with additional input data for specific city, state, or utility territories. ResStock is free, open source, and publicly available, with the large-scale simulations running on Amazon cloud computing. See resstock.nrel.gov for more information on using the software, and to learn how to partner with NREL or third-party consultants on additional analyses.

This improvement involves adding densely packed cellulose or fiberglass insulation to existing wood-framed wall cavities that are empty. Holes for adding insulation are drilled in each wall cavity (every 16 inches). This can be done from the outside, when it is convenient to remove a row of siding, or from the inside, which requires patching the holes made in the drywall or plaster.

References

• https://sws.nrel.gov/spec/411

This improvement involves adding R-5 of rigid foam sheathing (e.g., 1-in. rigid extruded polystyrene foam) or an R-5 insulated siding product at the time of residing. Thus, the cost of this improvement only includes the additional material and labor costs associated with the insulation; the removal of old siding and installation of new siding is not included.

The results of this analysis assume a siding replacement rate that ranges from 10% of 1980s homes to 50% of pre-1950s homes (see Wilson et al. 2017 for details).

References

• https://basc.pnnl.gov/resource-guides/continuous-rigid-insulation-sheathingsiding
• https://basc.pnnl.gov/resource-guides/rigid-foam-insulation-existing-exterior-walls
• https://sws.nrel.gov/spec/411033

This foundation insulation improvement adds R-10 (2 in.) of rigid extruded polystyrene foam to the interior side of foundation walls and rim joists in finished and unfinished basements. For this analysis, we assume that finished basements are heated and cooled, whereas unfinished basements are not directly heated or cooled.

While fiberglass batt insulation is sometimes used to insulate foundation walls, rigid foam board is considered best practice due to its superior durability when exposed to water. Note that local codes may require more than R-10 when insulating foundation walls.

References

• https://basc.pnnl.gov/resource-guides/unvented-crawlspaces-and-conditioned-basements
• https://sws.nrel.gov/spec/41402

This improvement involves installing a smart thermostat to control a home's central heating and/or cooling system. For this analysis, smart thermostats were modeled in a way analogous to programmable thermostats rather than attempting to model the specific control algorithms of one or more brands of smart thermostat and the associated occupant behavior patterns that would drive those control algorithms.

A study of 1,420 customers who installed one brand of smart thermostat found that the average nighttime setback was around 4 °F. With this in mind, a 4 °F thermostat setback in heating mode and a 4 °F thermostat setup in cooling mode was used to model the smart thermostat in the subhourly building energy simulations. We confirmed that this degree of setback/setup resulted in a similar range of heating and cooling energy savings as observed in field studies.

References

• https://sws.nrel.gov/spec/531041

This improvement brings the R-value of attic floor insulation up to R-60. Blown-in fiberglass and blown-in cellulose insulation have similar costs and performance, so a distinction is not made for this analysis.

Blown-in attic insulation only applies to vented attics, not finished attics or cathedral ceilings. This improvement is applied to all homes with vented attics with existing insulation that is less than R-60. R-38 and R-49 options were also considered, but this R-60 option had the largest economic potential for this state.

References

• https://basc.pnnl.gov/resource-guides/blown-insulation-existing-vented-attic
• https://sws.nrel.gov/spec/410

This improvement achieves a 25% reduction in building enclosure infiltration, as measured by a blower door test in units of air changes per hour at 50 pascals (ACH50). For improvements resulting in measured infiltration of less than 7.0 ACH50 (maximum allowed by the 2009 International Energy Conservation Code; roughly equivalent to 0.35 natural ACH), mechanical ventilation in the form of a bathroom exhaust fan operating continuously with flow rate specified by ASHRAE Standard 62.2-2010 is added to maintain indoor air quality. The air-sealing improvement applies to essentially all homes.

The value of 25% reduction was chosen based on a large-scale analysis of 23,000 before and after retrofit leakage measurements that found air sealing typically achieved a reduction in the 20%-30% range, with a median reduction of 25%. A 25% reduction was also used for a savings analysis for the "Seal and Insulate with ENERGY STAR" program. Reductions of more than 25% are certainly possible, but they may require more aggressive air-sealing improvements that are uncommon or involve other enclosureimprovements, such as R-5 wall sheathing insulation.

Other improvements, such as window and insulation retrofits, can also reduce air infiltration. These reductions are not included in this analysis.

References

• https://basc.pnnl.gov/search/air-sealing-guide
• https://sws.nrel.gov/spec/3

This improvement installs low-E storm windows on single-pane primary windows or double-pane primary windows. Homeowner self-installation was assumed to cost $8.30/ft^2. See Wilson et al. 2017 for details of performance (U-value and solar heat gain coefficient).

References

• https://sws.nrel.gov/spec/312

This improvement involves replacing 95% of the lamps (commonly called bulbs) in every home with high-efficacy light-emitting diode (LED) lamps (80 lumens per watt). A value of 95% replacement was assumed instead of 100% because of diminishing returns on replacing infrequently used lamps (e.g., in closets or storage areas). These diminishing returns are accounted for using the smart replacement algorithm in the Building America HSP.

This improvement is made today rather than upon wear out of existing lamps. The reference is to make no change. The distribution of existing lighting types was based on RECS 2009 microdata using the survey responses for "number of energy-efficient bulbs." The values queried from RECS are shown in Table 21. These values are consistent with recent numbers for market penetration of ENERGY STAR certified lamps. All existing "energy-efficient" lamps were assumed to be compact fluorescent lamps with an efficacy of 55 lumens per watt. The remaining lamps were assumed to be incandescent with an efficacy of 15 lumens per watt. The installed cost of LED lighting was assumed to be $0.12/ft^2 of living space.

References

• https://sws.nrel.gov/spec/780031

This improvement involves installing a new high-efficiency variable-speed heat pump (SEER 22, HSPF 10) upon failure of the existing centrally ducted oil furnace. This inverter-driven heat pump is representative of ducted VSHPs currently on the market. It meets NEEP's "Cold Climate Air-Source Heat Pump Specification," so it can be considered a cold climate heat pump (CCHP). However, unlike some newer ducted CCHPs, the representative unit does not retain heat pump capacity at temperatures below 0 °F, so the scenario includes an electric resistance coil for backup heat in parts of the United States with winter temperatures reaching near or below 0 °F. An alternative would be to retain the existing fuel-fired furnace for backup heat, but that scenario is not considered in this analysis.

The reference for these improvements is based on the current federal standard for furnaces of 82% for oil. Both the heat pump and reference replacement furnace capacities are sized in accordance with ANSI/ACCA Manuals J and S, except that the heat pump is sized based on the larger of the heating or cooling loads ("cold climate sizing") rather than the cooling load priority specified by Manual S.

Heat pumps provide both heating and cooling, so in homes that have existing air conditioning,these improvements provide cooling energy savings in addition to the heating energy savings. In homes without existing air conditioning, total energy use could increase because of new AC usage. These improvements were assumed to be triggered by failure of the existing furnace; if instead they were triggered by failure of a home's AC, the economics would be similar or better.

This improvement involves sealing and insulating any HVAC supply and return ductwork that is outside of conditioned space. For this analysis, duct leakage is defined as the fraction of air handler fan flow rate that leaks out of pressurized supply ducts into unconditioned space, plus the fraction of fan flow rate that leaks into depressurized return ducts from unconditioned space, plus any leakage into or out of the air handler itself (all measured at 25 pascals pressure difference).

The Building Performance Institute, which sets quality assurance and technical standards for home performance professionals, recommends 10% as the maximum allowable duct leakage. The Lawrence Berkeley National Laboratory's "Best Practices Guide for Residential HVAC Retrofits" also recommends a target duct leakage of less than 10% of air handler flow for duct sealing retrofits. Thus, 10% was used as the post-retrofit duct leakage value for this analysis.

This improvement also includes adding R-8 insulation to any uninsulated ducts located in unconditioned space. Ducts with existing insulation (typically R-4, R-6, or R-8) do not have any insulation added.

Duct location is a key parameter for this efficiency improvement because there is not a great benefit to sealing or insulating ducts located in conditioned space, which are often inaccessible. Duct and air handler location are not fields available from RECS, so we infer it from foundation type. Based on the Building America HSP, homes with slab foundations have their ducts in the attic, homes with crawlspaces have their ducts in the crawlspaces, and homes with basements have their ducts in basements (conditioned or unconditioned). Homes with two or more stories are assumed to have 35% of their supply ducts within conditioned space. Supply and return duct surface area is also specified as a function of finished floor area and number of stories.

There is a lack of good data on the range of duct leakage in existing homes, but one field study found an average of 21% leakage. For this analysis, we used a distribution of leakage based on a normal distribution centered at 20% leakage with a standard deviation of 8% leakage.

References

• https://sws.nrel.gov/spec/316
• https://sws.nrel.gov/spec/416