Appliance efficiency improvements

This post is the third in a series of posts exploring why residential fossil energy use in Massachusetts declined so dramatically even before the Green Communities Act of 2008 launched intensive energy efficiency efforts. Market forces together with federal laws and regulations have driven continuous national improvement in heating and other appliance performance since the 1970s. The evolution of federal regulation has a tortured history involving both litigation and directional shifts as administrations have changed. However, as of today, DOE appliance and equipment efficiency standards cover 90% of the energy use in homes.

The take away from this is post is that fossil appliance efficiency improvements were an important contributor to the 58% drop in residential fossil energy use intensity in Massachusetts from 1972 through 2009. The table below summarizes the estimates developed in the text further below.

Fossil Appliance Efficiency Improvement in Massachusetts, 1972 to 2009	Low Estimate	High Estimate
Space heating	18%	47%
Water heating and other appliances	9%	47%
All fossil appliances (weighted by 2009 energy use, 73.6% heating)	16%	47%
If no efficiency improvement, Fossil EUI in 2009 (57.1 kbtu/sqft/year) would be higher by:	9.2	27.1

Furnace and boiler improvements

How did the average annual fuel utilization efficiency for residential heating systems change in Massachusetts from 1972 to 2009?

Annual fuel utilization efficiency

AFUE is the ratio of the furnace’s or boiler’s annual heat output compared to its total annual fossil fuel energy consumed. An AFUE of 90% means that 90% of the energy in the fuel becomes heat for the home and the other 10% escapes up the chimney and elsewhere. AFUE doesn’t include the heat losses of the duct system or piping, which can be as much as 35% of the energy for output of the furnace when ducts are located in the attic, garage, or other partially conditioned or unconditioned space.

Energy.gov, Understanding the Efficiency Rating of Furnaces and Boilers

The AFUE numbers we see published are laboratory rating results. Real world performance is much more complex to measure. One of the early methodological studies for AFUE measures summarized the difference between the laboratory and the real world well:

The seasonal performance of . . . a device is not determinable on the basis of laboratory measurements alone. Seasonal performance is ultimately measured, by the consumer, in terms of operating costs for the installed device, serving a specific house, in a specific city, during a specific heating season, subject to the infinitude of house variables (house size, shape, degree of insulation, wind exposure, solar exposure, airtightness, internal heat sources, venting, sight topography, landscaping effects, etc.) as well as variables pertaining to the wisdom of the heating systems’ installation (line losses, design firing rate, use/nonuse of conditioned air for combustion, etc.) and servicing (cleanliness of boiler, fuel nozzle cleanliness, etc.).

Seasonal performance and energy costs of oil or gas-fired boilers and furnaces, Brookhaven National Laboratory, 1977, page 2.

Another study from the same source said:

The list of possible schemes by which one may degrade efficiency performance of an individual building is almost without limit.

Annual Fuel Use and Efficiency Reference Manual, Hydronic Equipment, Brookhaven National Laboratory 1977.

A variable of particular importance is the sizing of the heating system measured in maximum BTU per hour. Systems that are oversized for the home that they are serving will cycle on an and off more and so have more energy losses through idling, start up, and cool down. The oversizing factor is the percentage by which the capacity of the heating system exceeds the estimated maximum demand at design temperature, where design temperature is the coldest outdoor temperature encountered in the climate zone.

Methods for the standardized seasonal measurement of fuel efficiency were developed at Brookhaven National Laboratory in the mid 70s, and AFUE measurement standards were finalized in regulations by the federal Department of Energy in 1978 — see 43 Federal Register 91, 20147. The question of what oversizing factor to assume in laboratory AFUE testing was discussed in the DOE 1978 final rulemaking:

[S]ome commenters criticized the assumed oversizing factor of 70 percent used in the calculation of burner operating hours. In the field, many furnaces and vented heaters are considerably more than 70 percent oversized. Consequently, they operate fewer hours, at a higher firing rate, than would less oversized furnaces or vented heaters. However, the oversizing practices of the past are being moderated today because of an increasing awareness that oversized equipment is less efficient. Therefore, DOE believes the 70 percent oversizing factor is currently an appropriate factor to use in these calculation procedures.

43 Federal Register 91, 20147 at 20152, 20163 (1978)

The oversizing factor of 70% has continued to be used in AFUE calculations. In its 2015 rule making proposal to update consumer furnace and boiler test procedures, the DOE gave considerable attention to the possibility of updating the factor. In its final rule updating test procedures, DOE determined that

[T]he existing value of 0.7 continues to be representative of the oversized factor applicable to the average U.S. household.

81 Federal Register 2628, (2016)

In the 2023 rule making specific to boilers as opposed to furnaces, DOE again adopted a fixed 70 percent oversize factor, noting the lack of any solid data as to changes in field sizing practices.

Average fossil heating AFUE in 1972 in Massachusetts: 55% to 65%

We do not have a contemporaneous survey from which we can confidently derive an average AFUE of existing heating systems in Massachusetts in 1972, but a range of 55% to 65% seems reasonable.

As Brookhaven National Laboratory refined fuel efficiency measurement standards in the late 70s, they published various observations of existing installed boilers:

• Boiler with measured “Seasonal Efficiency” of 56% to 61% (when oversized 100%) deemed to be representative of field surveys of boilers. Direct measurement of the overall efficiency and annual fuel consumption of residential oil-fired boilers Annual report, Fiscal Year 1977. Brookhaven National Laboratory, 1978. See pages 6 and 79.
• “Seasonal performance factor” ranges of 67% to 81%, 53% to 72%, and 51% to 67% for three specific devices in the New York metropolitan area under alternative assumptions for oversizing and shared domestic hot water load. The low-end estimates reflect installations where the equipment is oversized by a factor of three (oversizing factor of 200%) and the high end estimates reflect an unrealistic oversizing factor of 0%. Seasonal Performance and Energy Costs of Oil or Gas Fired Boilers and Furnaces. Brookhaven National Laboratory, 1977.
• “BNL Annual Fuel Use Efficiency” ranges of 44 to 65%, 44% to 66%, for two specific devices in the New York metropolitan area under alternative assumptions for burner sizing relative to heating load and shared domestic hot water load. The low-end estimates reflect installations where the equipment is oversized by 300%. Oversizing by 100%, deemed “not uncommon” in the field yielded AFUEs in the high 50s. Direct efficiency measurement and analysis of residential oil-fired boiler systems, Brookhaven National Laboratory 1977.
• For additional similar results, see Annual Fuel Use and Efficiency Reference Manual, Hydronic Equipment, Brookhaven National Laboratory 1977.

From the early 80s, we find some additional retrospective data points.

• Average AFUE measurements for heating systems shipped in 1978: Indoor gas furnace, 65%; indoor gas boiler, 65%; indoor oil furnace, 75%; indoor oil boiler, 76%. 45 Federal Register 127, 44002 (1980).
• “Baseline: Typical of Conventional [Oil] Burner Boilers . . . . Seasonal Efficiency Range . . . 63-67%”, Technology for the Development of High-Efficiency Oil-Fired Residential Equipment, page 32, Table III-1, Brookhaven National Laboratory, 1980.
• “Traditional” indoor gas-fired furnaces and boilers, AFUE of 65%; “traditional” oil-fired furnaces and boilers, AFUE of 63-71%. Efficiency Improvements in Space Heating by Gas and Oil, Annual Energy Review 1983, 8:247-67. This is a useful broad review of heating technology at the time.

Finally three current respected sources broadly use 65% as an AFUE estimate for older heating systems:

• An Energy Information Administration study states (in unsourced Footnote10): “A standard gas furnace in the 1970s had an annual fuel utilization efficiency (AFUE) of 65%.”
• Consumer Reports states without citation “A gas furnace made in the early 1970s typically has an AFUE rating of about 65 percent.”
• Another DOE page characterizes “Older, low-efficiency heating systems” as having “56 to 70% AFUE.”

To these general national sources, we add two observations from local experts:

• Mark Dyen, an energy efficiency professional active in home energy audits in Massachusetts in the 70s and 80s, recalls 65% as a common working assumption for Massachusetts existing seasonal burner efficiency.
• Charles Uglietto, a Massachusetts home oil business executive, recalls that many of the systems that his teams were replacing in the 70s were older boilers previously converted from coal and were very inefficient — in the low 50s. This is consistent with historical fuel use data presented in this spreadsheet, tab E-Heat-Prev, which shows coal declining from 57% of Massachusetts homes in 1940 to 35% in 1950 to 7% in 1960 to 0% in 1970. Two thirds of Massachusetts homes were heated by oil in 1970.

Several statements above characterize average AFUE as about 65% in the 1970s. It’s not always clear whether commenters are referring to systems shipping in the 1970s or to the installed base in the 1970s and none of the commenters can speak from rigorous historical survey data. It appears from Mr. Uglietto’s experience that the oldest, most inefficient installed heating systems in Massachusetts were running below 55% AFUE. Today, Massachusetts, as a colder state, might be expected to have more efficient systems than some other states, but before the energy shocks of 1973, technology development for heating systems was focused on cost and reliability (see Technology for the development of high-efficiency oil-fired residential heating equipment. Final report, Brookhaven National Laboratory, 1980). And since Massachusetts has an older housing stock than some other parts of the country, the population of heating systems might have been older and less efficient than the national average. However, even in Massachusetts many of the systems in place in 1972 had to be newer and more efficient than the oldest coal-converted systems.

It appears that the most we can say as to average installed Massachusetts AFUEs in the early 1970s is that it was likely in the range of 55% to 65%.

Average fossil heating AFUE in 2009 in Massachusetts: 77% to 81%

We have somewhat more useful estimates relevant to average Massachusetts installed AFUE around 2009. All are consistent with a range of 77% to 81% AFUE for the installed base.

• AFUEs between 75% to 80% were used in 2009 to characterize systems to be replaced by high gas efficiency burners in Mass Save incentive program. 2009 Report on Natural Gas Energy Efficiency Potential in Massachusetts, Table A-2B. This does not necessarily represent the existing installed base, it could represent only the burners likely to be replaced within the installed base — the report is not clear.
• From a national DOE study on Buildings Sector Appliance and Equipment Costs and Efficiencies, we have the following AFUE estimates for the national installed base for 2015.
  • Gas-fired furnaces (Northern), 80% (same as nation-wide)
  • Oil-fired furnaces, 83%
  • Gas-fired boilers, 82%
  • Oil-fired boilers, 84%
• From Mass Save’s Residential Baseline study, we have the following AFUE estimates for averages in 2019 (computed from binned data in the attached spreadsheet)
  • Gas-fired furnace, 86%
  • Gas-fired boiler, 85%
  • Oil-fired boiler, 84%

The appendices to the 2019 baseline study of Massachusetts homes include estimates of the “saturation” or prevalence) of different types of heating equipment, as well as, for the more prevalent types, the characteristics, including AFUE. Together these data yield 85% as an estimate of the average rated AFUE of existing fossil-fuel space heating systems in Massachusetts in 2019. The 2019 Baseline data is the only actual Massachusetts data we have, but we should take it as a high upper bound on the efficiency in 2009. Mass Save was offering incentives for heating system efficiency upgrades during many of the intervening ten years.

Replacement rates for furnaces and boilers run in the range of 3% to 6% per year:

• From the Residential Energy Consumption Survey, in 2020, we know that only 24% of main heating equipment in the “Cold” climate zone which includes Massachusetts was over 20 years of age, indicating that 76% had been replaced in the past 20 years — a 3.8% replacement rate.
• From the Residential Baseline study, Figure 3-10, we have equipment mortality/replacement rates in Massachusetts ranging from 3% for oil boilers to 6% for gas furnaces.

At 4% per year, 40% of the installed heating systems would have been replaced from 2009 to the 2019 Baseline Study. Many of those systems had already been replaced once since 1972, but most of those upgrades likely involved significant upgrades in efficiency. If those upgrades were mostly to the oldest equipment and averaged 10 AFUE to 20 AFUE points, the impact on the base average AFUE would be 4 to 8 AFUE points. Working backwards from the 2019 figure of 85%, it follows that 77% to 81% is a reasonable range for the installed base AFUE in Massachusetts in 2009 — a range roughly consistent with the other data sources above.

For one more data point broadly consistent with the 77% to 81% range, we can look at shipments from 1992 to 2009 and combine them with earlier shipment data to work forward to our range. In 2017, Navigant (recently acquired by Guidehouse) did a study for DOE attempting to quantify the efficiency of shipments of different appliances through the years.

Heating System Type	1992 Minimum Efficiency Rule	Available data on shipment trends before 2009	2019* Saturation
Gas furnace	78%	from 2004 to 2009, a little over half of natural gas furnaces shipped with AFUEs at 78% to 80%, with the rest over 90%	24%
Gas boiler	75% for steam; 80% for hot water	from 2003 to 2009, over half of natural gas boilers shipped with AFUEs under 84%, but the over 90% category exceeded 1/3 of shipments by 2008	33%
Oil furnace	78%	from 2004 to 2009, most oil furnaces shipped with AFUE in the 78% to 84% range	4%
Oil boiler	80%	from 2003 to 2009, over half of oil boilers shipped with AFUE of over 85%	17%

At 4% per year, perhaps two thirds of Massachusetts heating units were replaced during the 17 years from when Congress implemented minimum efficiency standards in 1992 to our subject 2009. As shown in the chart above, most types of units shipped in this period had AFUEs of 78% or better by law. However, as noted further above, shipments in 1978 averaged roughly 70% AFUE (or lower since oil units of that era were more efficient at 75% than gas at 65% but gas was displacing oil in Massachusetts). If in 2009, the oldest 1/3 of Massachusetts units still had an average AFUE efficiency of 70% and the 2/3 more recently shipped to Massachusetts had an average AFUE of 85%, averaging well above the statutory minimum, the weighted average AFUE would be 80%, within our estimate range of 77% to 81%. Reasonable variations of this computation produce similar results.

Improvement in average fossil heating AFUE in Massachusetts, 1972-2009: 18% to 47%

Our low estimate of AFUE improvement, 18%, is the improvement based on the high estimate AFUE for 1972 and the low estimate for AFUE in 2009. The high estimate, 47%, is the reverse. In summary, our estimate is that fossil heating load would be at least 18% higher in 2009 if there had been no AFUE improvement, and at most 47% higher.

Water Heating and other Appliances

Oak Ridge National Laboratory appears to have done the early work on water heater efficiency measurement. Their 1978 publication, “An efficiency evaluation and consumer economic analysis of domestic water heaters,” highlights the variables one must specify and hold constant to compare the performance of water heaters, including water usage levels and patterns which vary widely in the field. Water heaters differ from space heaters in that a high proportion of down-time is to be expected in every season; almost half of hot water energy is wasted as people run faucets to purge cold water, although this factor does not enter the definitions of performance metrics.

While, in basic load concept, the AFUE for spacing heating has remained stable, the performance metrics for water heating appear to have evolved more significantly. The original rule promulgated in 1977, 42 CFR 54110 was soon modified soon after in 1979, 44 FR 52632, 1979. The currently effective rule uses a much more complex test procedure than the original. It handles a variety of water heater variations and simulates water use patterns in a more detailed way. The corresponding measurement formulas are more complex. Many test specific parameters differ, including tank thermostat settings, input water settings, ambient air temperature. DOE’s recent final rule updating water heater testing standards will take effect December 2023.

Perhaps because of the complexity and the changing concepts, but also perhaps because of the smaller overall energy draw, the historical measurement literature on water heating seems much thinner than the literature on space heating. An additional reason for the older literature on water heaters per se being thinner is that water heating was often integrated with space heating — all the old Brookhaven studies cited above involved boilers that supplied hot water for domestic use as well as steam or hot water for the heating system. New water heater designs allowed separation of hot water for domestic use from hot water for heating from boilers and relieved boilers of the need to keep a large amount of water hot even in the summer. This was a big contributor to the improved efficiency of boilers discussed above.

The old boilers that also supplied hot water were in the 55% to 65% AFUE range. While water heating clearly dragged down boiler operating efficiency, it is conceptually difficult to assign a distinct AFUE to the water heating component of boiler load. Historical water heating in boilers should be thought of as having the same efficiency as the overall efficiency of the boiler — 55% to 65% AFUE. Today, the intrinsic inefficiency of water heating, with its surge and standby requirements, is now isolated and visible in a dedicated and much smaller appliance. And, in fact, modern Uniform Energy Factors specifically for common water heaters still run in the 50% to 60% range. Arguably, the efficiency of water heating did not improve in the study period for the share of homes that used boilers. We have do have a data point on boiler prevalence from the 1981 Residential Energy Consumption Survey: A little over half of homes in the Northeast had boilers at that time. Surely, the dedicated water heaters that might previously have been used in homes with furnaces have improved considerably — in some of the same ways as furnaces, and in other ways — but we lack good data to characterize their improvements.

In the absence of better data we make the following alternative assumptions as to the improvement of efficiency of hot water heating systems in our study period: (1) Half did not improve; half improved as much as space heating systems in the lower improvement scenario. (2) All improved as much as space heating systems did in the higher improvement scenario.

Other Appliances

Gas dryers and kitchen equipment have improved during the study period through the transition from pilot lights to electronic ignition and undoubtedly in other ways. What those improvements have added up to is probably impossible to know. Fortunately, these appliances account for a small share of use, 8% in Massachusetts according to the 2009 RECS data. For simplicity, we use the same alternative improvement assumptions as for hot water systems.

Resources

• Computations directly supporting this post
• Energy use data, including breakdown of energy use in 2009
• Historical Energy Use Intensity computations

Next or Return to Series Intro or Return to heat pump outline