This article was originally published in the May/June 1995 issue of Home Energy Magazine. Some formatting inconsistencies may be evident in older archive content.



| Back to Contents Page | Home Energy Index | About Home Energy |
| Home Energy Home Page | Back Issues of Home Energy |



Home Energy Magazine Online May/June 1995

Measuring the Performance of the
National Energy Audit

by Terry Sharp

A field test conducted in North Carolina demonstrates significant performance improvements that the National Energy Audit (NEAT) offers compared with Project Retro-Tech, and helps show that state-of-the-art weatherization methods have come a long way.

In 1978, the U.S. Department of Energy (DOE) published Project Retro-Tech to provide all states a manual technique for identifying low-income weatherization measures that would produce the most energy savings per dollar spent. While Project Retro-Tech was considered a significant advancement at the time, it focused only on a small number of shell measures and relied upon broad assumptions and limited analytical capabilities because of its manual calculations. In addition, these manual calculation requirements were challenging and time consuming for auditors.

During the 1980s, a few states began to incorporate promising new technologies into their weatherization programs. Many of these new additions offered dramatic performance improvements over Retro-Tech alone. Simultaneously, two large demonstration tests in Wisconsin and New York incorporating advanced measures, techniques, and evaluation methods showed that weatherization performance could be doubled without significantly increasing costs. Federal regulations are now encouraging states to use more advanced weatherization technologies. An advanced measure-selection method developed by Oak Ridge National Laboratory and currently being used by many states, is the National Energy Audit--NEAT (see Computerized Energy Audits, HE May/June '94 p.27).

The transition from the preliminary to the current version of NEAT required expansion to national application and many additions to meet the needs of potential users, but there were few changes to its design fundamentals. Between 1989 and 1991 we tested the preliminary version of NEAT alongside North Carolina's Retro-Tech-based weatherization program. The cooperative test involved DOE, the State of North Carolina, the Alliance to Save Energy, and three local weatherization agencies.

Retro-Tech-Based Weatherizations

In 1989, North Carolina's Retro-Tech-based program was limited to the following measures (sometimes called the big six) in order of their installation priority:

  • Infiltration measures
  • Attic insulation
  • Water heater, pipe, and floor insulation
  • Duct insulation
  • Underpinning (enclose the crawl space)
  • Storm windows and storm doors

Each measure was installed until the allotted funds, up to $1,400 per house on average (including administration costs), were expended or the next consecutive measure was deemed unaffordable. Needed infiltration (air sealing) measures were identified by the auditor and included caulking, adding or replacing weatherstripping, repairing or replacing windows and doors, and other repairs. Attic insulation was installed to a minimum of R-19 and a maximum of R-30. R-11 or R-19 floor insulation was installed.

The Advanced Field Test Audit (FTA)

In contrast to Project Retro-Tech, NEAT is a computer-based audit that addresses many shell and equipment measures aimed at reducing both space heating and cooling energy. Instead of a set list of priorities, NEAT evaluates the expected performance of each measure according to the needs of the individual house. The program ranks measures by their benefit-to-cost ratio (BCR), adjusts this ranking for interactions between measures (very important where both mechanical and shell measures are installed), and provides a final list of recommended measures with a benefit-to-cost ratio greater than one. The program quantifies costs and savings, and recommends measures within seconds.

NEAT considers local costs (labor, materials, and fuel), measure savings (both heating and cooling), measure life, and the current discount rate. Local costs, house descriptive details (size and construction), equipment details, and results from diagnostic measurements (air leakage and heating system efficiency)--if made--are provided as input by the auditor. The preliminary version of NEAT, the advanced Field Test Audit (FTA) used in the North Carolina study, contained all of these features.

Air sealing in FTA-treated houses was done independent of the FTA using a blower-door-directed procedure and a cost-effectiveness guideline ( see Blower Door Guidelines for Cost-Effective Air Sealing,HE Mar/Apr '90 p.34). A cutoff of 75 CFM50 (cubic feet per minute at 50 Pascals house depressurization) air leakage reduction per person-hour of labor was used for cost effectiveness and a minimum ventilation guideline of 1,500 CFM50 was used for sealing.

Auditors also made decisions about heating system tune-ups outside of the FTA, based upon steady-state efficiencies determined from actual flue gas measurements on each system. (These decisions can now be done within NEAT.)

Figure 1. Installed measures with frequencies greater than or equal to 10% in either group.

Field Test Design

The test began with 120 houses at three weatherization agencies (sites). Houses were heated by kerosene, fuel oil, natural gas, or propane, and had one or two operating window air conditioners. Houses were split into three groups of 40 representing FTA, Retro-Tech, and control groups, with each of the three test sites represented approximately equally (13 or 14 houses per site in each group). We monitored weekly space heating and cooling energy use and hourly indoor and outdoor temperatures between November 1989 and September 1990 (pre-period) and between December 1990 and August 1991 (post-period).

We created linear models of measured heating and cooling energy consumption as a function of average indoor-outdoor temperature difference and weather-normalized them to adjust for differences in average seasonal temperatures between the two monitoring periods. We determined energy savings from the differences between pre- and post-weatherization normalized consumptions.

Installed Measures

Of the 24 measures considered in the two weatherization procedures, Retro-Tech considered 10, while FTA weatherizations considered 21 (see Table 1). Of these 21, the FTA made the installation decision for 17. The FTA considered shell and mechanical equipment measures that the Retro-Tech audit did not.

Attic, wall, and floor insulations dominated the FTA-recommended measures and were installed in 40%-70% of all houses in this group. Attic and floor insulations dominated Retro-Tech-recommended measures in quantities similar to the FTA group (when both floor R-19 and R-11 insulations are combined). The largest differences were for wall insulation and storm windows. Retro-Tech did not install wall insulation (because it was not an option), but the FTA recommended it nearly half of the time. In addition, Retro-Tech weatherizations resulted in new storm windows on more than 80% of all houses, while less than 5% received storm windows in the FTA group (due to poor window condition--not FTA-recommended).

The FTA installed R-30 in all uninsulated attics and either none, R-11, or R-30 in those with insulation (the FTA did not consider R-19 at the time). The Retro-Tech procedure filled all attics to an R-30 level, independent of the existing attic insulation level. Except for infiltration reduction, attic, wall, and floor insulation, and storm window measures, neither audit called for the installation of any other measure in greater than 11% of their respective houses.

The FTA recommended very few heating system measures. This was primarily due to the abundance of space heaters (used for primary heating in nearly 70% of all houses), which were not considered compatible with most heating system measures. A replacement air conditioner was recommended for one house (3%), but this house dropped out of the test before it was weatherized.

The percentages of FTA houses receiving some measures were very high based on the number of houses that could actually have the measures installed (only houses with central furnaces for smart thermostats, with accessible wall cavities and floors for wall and floor insulation, and so forth). Attic, wall, and floor insulations were recommended for more than 80%, kneewall insulation was always recommended, and smart thermostats were recommended for 57%.

Air Sealing Results

We estimated air leakage rates at 50 Pa depressurization in all houses before and after weatherization. Estimates were determined from multiple blower-door measurements and a regression-based estimation procedure similar to that specified by ASTM Standard E779-81. The best linear fit of air leakage rates measured at five different house depressurizations between 20 and 60 Pa was used to estimate house air leakage rate at 50 Pa. For high-leakage houses, those that could not be depressurized to 50 Pa, air leakage rates at 50 Pa were extrapolated from regression models based on five measurements made below 50 Pa.

The average pre-weatherization air leakage rate for all groups was 4,300 CFM50. Individual group averages were within 5% of this value. Post-weatherization measurements were not completed in the control group at two sites. As a result, reported air leakage reductions represent 12 houses in the control group, 36 in the FTA group, and 37 in the Retro-Tech group. The average air leakage reduction was 89 CFM50 for the control group (no treatment), 1,710 CFM50 for the FTA group, and 716 CFM50 for the Retro-Tech group. Adjusting for the control group, reductions of 37% were achieved in the FTA group, compared to 16% in the Retro-Tech group. The average leakage reduction of the control group was within the expected error of blower door test measurement.

Space Heating

Sixteen of the original 120 houses were removed from the heating energy savings analysis due to attrition and other data problems. We chose 65 of the remaining houses for a heating analysis, after using a statistical analysis to screen out those with unreliable data (see Now That I've Run PRISM, What Do I Do with the Results, HE Sept/Oct '90, p. 27). Generally, unreliable models occurred where the consumption data were highly scattered (nine or fewer data points available for either period) or represented only a small range of expected seasonal temperatures.

Weather-normalized, pre-weatherization space-heating energy consumption averaged 50 MBtu (around $350 at $7/MBtu) for all 65 houses, and the three group averages were within 3% of this value. Space-heating energy savings occurred for 89% of the FTA group and 87% of the Retro-Tech group.

Individual space-heating savings of the 65 houses ranged from a low of -33 MBtu (an increase in energy use) to a high of 53 MBtu. Average savings for control houses was -2.7 MBtu (-5%). Average energy savings for the two weatherized groups were much larger, 13.9 MBtu (28%) for the FTA group and 8.9 MBtu (18%) for the Retro-Tech group (see Table 2). Adjusted for the control group, FTA weatherizations saved 33% and Retro-Tech weatherizations saved 23%. This represents a performance improvement of 43% over Retro-Tech-based weatherizations.

Space Cooling

Twenty houses were removed from the air conditioning sample due to attrition and as in the heating energy analysis, many cooling energy models were found to be unreliable. Houses with unreliable models were not removed from the cooling analysis, however, because the lower reliabilities resulted from the minimal and random use of air conditioning in North Carolina's low-income houses--not data problems.

Weather-normalized, pre-weatherization space-cooling energy consumption ranged from 0-4,867 kWh. The average pre-weatherization cooling energy use for the 100 houses was 781 kWh (around $66 at $0.085/kWh); the three group averages were all within 13% of this value. Most houses in the test were very low cooling energy users. Approximately one of every two houses (47%) used less than 500 kWh ($43) before weatherization. Only 14% used more than 1,500 kWh ($128).

The average cooling energy savings was 8 kWh (1%) for the control group, 30 kWh (4%) for the Retro-Tech group, and 165 kWh (19%) for the FTA group. When adjusted for the control group change, the FTA group saved 18% (net savings), which is much higher than the 3% net savings for the Retro-Tech group. However, wide variances for the cooling energy savings of individual houses preclude saying there is a statistical difference between savings of the control and Retro-Tech groups.

Weatherization Costs

Although average weatherization costs for each group are almost identical ($1,056 versus $1,059) individual site averages are very different (see Table 3). Average FTA weatherization cost was 14% less at Site A, 7% less at Site B, and a dramatic 38% higher at Site C. Retro-Tech weatherizations at Site C also cost much less than Retro-Tech weatherizations at the other two sites (around 40% less). For Retro-Tech weatherizations, attic and floor insulation measures were much more common at Sites A and B than at Site C.

Individual house weatherization costs in the FTA group were more widely distributed. Approximately 90% of the Retro-Tech-based weatherizations cost between $500 and $1,500, while only about 70% of the FTA weatherization costs were within this range. Also, no Retro-Tech-based weatherization cost more than $2,000, while four houses (11%) exceeded this expenditure in the FTA group. The percent of weatherization dollars spent on labor costs (including labor for repairs and air sealing) was about the same, as shown in Figure 2. Thus, total material expenditures were similar, but the proportions spent on the different measures varied dramatically. Insulating materials in FTA houses accounted for about 34% of costs compared to only 18% in the Retro-Tech houses. Storm windows, which accounted for approximately one-third of all material costs in the Retro-Tech houses, were almost unrepresented in the FTA material costs (less than 2%).

Figure 2. Comparison of costs for the two weatherization approaches. (Note: costs other than the labor column are for materials only.)

In Table 4, we compared the performance of the two weatherization approaches to statisticsfrom past evaluations of standard weatherization programs and the three demonstration programs that represent the progression of NEAT technology. The performance of both standard and demonstration programs has been improving over time. The National Evaluation for 1989 indicates that standard program performance improved about 80% in eight years. In Wisconsin, however, weatherization performance had almost doubled in a demonstration program as early as 1985. More recent demonstrations in New York and North Carolina suggest that a standard program performance of 10% could be doubled and perhaps tripled by application of newer technologies. Also, while heating energy savings are increasing, average weatherization costs are decreasing. The 33% heating energy savings in North Carolina is particularly noteworthy because heating energy savings in warm climates are traditionally lower than the national average.

NEAT in '95

Today, NEAT uses weather data from 215 weather stations across the country to evaluate measures for 48 states (Alaska and Hawaii to be added as needed). While the fundamentals of how NEAT evaluates and prioritizes measures has seen only small change since the preliminary version, NEAT has undergone dramatic change based on needs indicated by current and potential users. Perhaps its most important area of change is in ease of use. Elements include recall of past building descriptions for editing or reuse; pop-up screens that ease data entry; duplicate, erase, and move features that allow building components to be rapidly described or changed; built-in capabilities to minimize user errors; defaults where measured values or equipment nameplate specifications are unavailable; and the ability for users to define their own measures.

Important equipment additions include the ability to address steam and hot water boilers, heat pumps, evaporative coolers, mandatory and optional heating system replacements, heating system tune-ups, and sizing for new heating systems. Other important capabilities include the ability to use billing data for comparison to or modification of NEAT results; to incorporate repair, health and safety, and miscellaneous costs into the weatherization package as appropriate; and to evaluate the energy savings and cost-effectiveness of air-sealing work.

NEAT is now a refined measure selection tool that can be easily tailored for local use. It addresses the wide range of measures and incorporates sophisticated analytical capabilities that are needed to advance weatherization performance nationwide. Hundreds of improvements have resulted from the recommendations of North Carolina users and others in states that have more recently evaluated NEAT for their programs.

The measure evaluation technology in NEAT pushes the most cost-effective of a large number of both heating and cooling measures to the highest installation priorities. As a result, it is adding new measures and excluding some measures that are traditionally installed in many state weatherization programs. In turn, changes in where dollars are spent and how they are distributed across houses are occurring.

Like other audits, NEAT is dependent on the well-trained auditor and does not solve all of the auditor's challenges. It is, however, a very valuable tool that can be incorporated into a state program to achieve substantial and perhaps dramatic improvements in weatherization performance. And most importantly, this can be accomplished without increasing costs.


M. A. Brown et al. National Impacts of the Weatherization Assistance Program in Single-Family and Small Multi-Family Dwellings, ORNL/CON-326. Oak Ridge, TN: Oak Ridge National Laboratory, May 1993.

McCold, L. N. et al. Field Test Evaluation of Conservation Retrofits of Low-Income, Single-Family Buildings in Wisconsin: Audit Field Test Implementation and Results, ORNL/CON-228/P2. Oak Ridge, TN: Oak Ridge National Laboratory, June 1988.

Sharp, T. R. The North Carolina Field Test: Field Performance of the Preliminary Version of an Advanced Weatherization Audit for the Department of Energy's Weatherization Assistance Program, ORNL/CON-362. Oak Ridge, TN: Oak Ridge National Laboratory, July 1993.

M. P. Ternes et al. The National Fuel End-Use Efficiency Field Test: Energy Savings and Performance of an Improved Energy Conservation Measure Selection Technique, ORNL/ CON-303. Oak Ridge, TN: Oak Ridge National Laboratory, January 1991.

Terry Sharp is a building energy researcher with Oak Ridge National Laboratory in Tennessee.
The publication of this article in Home Energy was underwritten in part by the U.S. Department of Energy's Office of Energy Efficiency and Renewable Energy.


  • 1
  • NEXT
  • LAST
Email Newsletter

Home Energy E-Newsletter

Sign up for our free monthly

Harness the power of

Get the Home Energy