Technical Description & Documentation

Following is a summary of the technical features of the Home Energy Saver. (Technical details on IT can be found here.) More in-depth engineering documentation is gathered here. The Home Energy Saver Team has also conducted an extensive review of web- and disk-based home energy calculators, including comparisons of features and an exploration of accuracy issues (Mills 2002 and 2004).

About the Heating and Cooling Module

The Home Energy Saver calculates heating and cooling consumption using the DOE-2 (version 2.1E), building simulation program (version 2.1E), developed by the U.S. Department of Energy. The program performs a full annual simulation for a typical weather year (involving 8760 hourly calculations -- using air temperatures, relative humidities, solar radiation, etc.) in about 2 seconds, after the user assembles the necessary information describing their home. Users can choose from almost 300 weather locations around the United States. DOE-2 performs a very sophisticated series of calculations, but the web-based user interface is relatively simple and results are distilled into a useful form. Calculations are done using a Building Description Language, or BDL, which consists of detailed information on the dimensions, materials, and construction of the foundation, walls, doors, windows, ceiling, and roof of a building, through which heat flows; the quantities and schedules of occupants, appliances, lighting, and equipment that generate heat inside the building; the sizes, efficiencies, and schedules of the heating and cooling systems used to condition the building; and, if desired, the local utility rates for the energy sources used to operate the heating and cooling systems (Birdsall et al. 1990).

DOE-2 is a complex physical model that accounts for the interactions between the weather conditions, building structure, and interior conditions. The first part of the DOE-2 program, the LOADS subprogram, determines the heating and cooling required in the building, given the weather conditions imposed on it. The SYSTEMS subprogram, the second part of DOE-2, calculates the energy used to operate the heating and cooling systems, and the degree to which these systems satisfy the comfort conditions required in the building. The third part of DOE-2, the PLANT subprogram, calculates the energy used to operate central heating and cooling systems, such as boilers and chillers, if such systems are used to feed individual heating and cooling units. Based on the specified utility prices, the ECONOMICS subprogram estimates the costs of the energy resources expended in operating the building conditioning systems.

About the Water Heating Module

To prepare the estimates, we use a detailed method developed here at Lawrence Berkeley National Laboratory. The original development of the water heating analytical method was sponsored by the U.S. Department of Energy, Office of Building Technology, State, and Community Programs as part of their appliance standards analysis program.

This module calculates energy consumption for heating water in three steps. The first step is to estimate average daily hot water use. This calculation is based on number and ages of people living in the house, presence or absence of a dishwasher and a clothes washer, the water heater temperature setting and tank size, and the local climate (Lutz et al. 1996).

Once the average daily hot water use has been estimated, a simple calculation is done to determine the daily energy use by the water heater. The calculation uses the energy consumption characteristics of the water heater as determined by the DOE Energy Factor test, air and water temperatures, and how much hot water is used on an average day. The last step is to convert the daily energy use into annual consumption of specific fuels, (e.g. electricity and gas).

Once energy use is calculated, we multiply by the energy price to arrive at the final energy cost. Prices for each fuel are defaulted to the average for the state in which the building is located (using the 1994 annual averages reported by the Energy Information Administration), or they can be overridden with custom inputs by the user.

You'll recognize some of the equation variables as things you are asked to provide on the web page. To simplify the process, the rest has been defaulted for you.

About the Appliance Module

The first step is to identify the appropriate unit energy consumption (UEC) for the technology in question. Large appliances, such as refrigerators, can have very different energy consumption depending on the year of manufacture, and features which affect energy use, like automatic defrost or side-by-side design. To estimate the energy consumption of these appliances, we utilized shipment weighted energy consumptions. This number is the average consumption for all units sold within a particular year weighted by the number of units in each consumption bin. This information was provided by AHAM, the Association of Home Appliance Manufacturers [AHAM 1996, and later years]. User inputs provide information about the appliance type and year. They are used in selecting energy consumption. For smaller appliances, and miscellaneous end-uses, a generalized annual UEC was used. This number is an estimate of the total energy consumed in a year, based on power draw and assumptions about typical usage. The specific assumptions pertaining to small appliances are based on the synthesis of many years of research here at Lawrence Berkeley National Laboratory (Sanchez 1997).

In situations where differing usage for an end-use results in large changes in energy consumption, consumers are encouraged to enter their indvidual hours of usage, which go into calculating the energy consumed. To hasten the inputting process, assumptions about usage were used for the majority of smaller end-uses.

Finally, default selections were set up to reduce the amount of input necessary to complete the module. Enduses having a saturation of 86% or greater in the overall housing stock are selected on the web page when it first loads (Sanchez 1997).

Once energy use is calculated, we multiply by the energy price to arrive at the final energy cost. Prices for each fuel are defaulted to the average for the state in which the building is located (using the 1994 annual averages reported by the Energy Information Administration), or they can be overridden with custom inputs by the user.

About the Lighting Module

For a high-level estimate lighting energy consumption, metered lighting data from the Tacoma Public Utility report was used (Tribwell and Lerman 1996). This report provides average lamp power (watts) and average usage by room. Users of the module are asked for the total number of lamps in each of a series of rooms. For each room the number of lamps is multiplied by the average wattage and hours of use to arrive at a typical energy consumption for that room. Obviously, house with non-typical usage, for example a light that is constantly on, or situations where the homeowner has already upgraded to efficient lighting technology will not be accurately modeled using the method described above. For these situations, we developed the detailed lighting module. The detailed module allows users to enter wattage and hours of use for every fixture in the house, thus providing a fully-customized calculation (Jennings et al. 1996).

About the Tariff Module

Once energy use is calculated, we multiply by the energy price to arrive at the final energy cost. Prices for each fuel are defaulted to the average for the state in which the building is located (using the annual averages reported by the Energy Information Administration), or they can be overridden with custom inputs by the user. While the calculations default to actual state-level electricity price, using real-world energy tariffs canmake the Home Energy Saver's results more realistic. The HES tariff methodology accommodates the following rate design features for electricity tariffs:

  • Fixed, energy and demand charges
  • Block rates with constant or variable block sizes
  • Hours charges, seasonal rates, time-of use rates

For more documentation on the tariff methodology, see the master documentation and the Tariff Analysis Project website and a special report.

About the Carbon Footprint Map

Emissions for the existing house are plotted on the map, for the address given. The placemarker is color-coded to indicate the amount of emissions, with the actual data shown when the marker is clicked. The primary "bell curve" distribution shown under the Carbon-IQ menu represents the entire population of Home Energy Saver users. These curves can be generated for any zip code.

To arrive at the carbon-dioxide (CO2) emissions for a given home, we multiply the annual energy for each fuel type by the CO2 emissions factor for the respective fuel. Electricity emissions factors are from U.S. EPA's eGRID (Emissions & Generation Resource Integrated Database), which contains emissions and resource mix data for virtually every power plant and company that generates electricity in the United States (US EPA 2009). Natural gas and fuel oil emission factors of 116.83 and 161.08 lbs CO2/million BTU are derived from U.S. DOE (1994), while the LPG emission factor of 137.26 pounds CO2/million BTU is from U.S. DOE (1996).

Default House Characteristics, Energy Consumption, and Savings

The Home Energy Saver utilizes default house characteristics, typical consumption, and savings values in order to generate approximate values for "typical homes" based on their location. The values are held in the database table CLIMATEZONES. Inputs (equipment and home characteristics), and consumption values are based on analysis of RECS data, primarily 2005 microdata [download microdata] updated with the 2009 public release tables. Because RECS content changes over time a few estimates (e.g. electrical cooking) are from previous surveys. Savings are based on existing estimates of the percent of savings achievable by upgrading to Energy Star or best available equipment.

Consumption and characteristics are also based on RECS. The analysis is based on a subset of the RECS homes, specifically mobile homes and single family homes (attached and unattached).

The original HES defaults were based on LBNL's 45 PEAR (Program for Energy Analysis of Residences) regions. Currently the 19 LBNL climate zones are used (Ritschard et al 1992, and Huang et al 1997, as cited in J. S. Apte "Residential windows, Greenhouse Gas Emissions and the Potential of Emerging Window technologies", May 2004 Baccalaureate thesis, Brown University). While these zones are numbered 1 through 20, there is no zone 4.


Heating fuel type is based on the RECS FUELHEAT variable (category of fuel used for space heating). Water heating type is based on the RECS variable FUELH20, category of fuel used for water heating. Both space heating and water heating fuels were re-coded to reduce variability when used as grouping variables. In the prior HES version, percentages of fuel were calcuated and the highest percent selected and as part of that LPG and NG were combined. In this version the modal values were obtained directly and no recoding of the RECS fuel types done. In determining the typical configuration inputs, the variable EQUIPM was used for heating equipment, as it accounts for heat pumps that way.

Air conditioning type is represented two ways: as Yes/No for central system; which is derived from RECS' COOLTYPE and CENACHP. If there is a central A/C with or without additional room units or a heat pump used for AC, the house is coded for central AC. The second typing is fourfold: No A/C, Central A/C present, Only Room A/C, and Heat Pump.

First cooling type was used to set presence/absence of cooling in the house for a particular climate zone. This was then used in conjunction with heating fuel and water heating fuel to select the subset of RECS houses that matched the "typical configuration package" to determine heating, water heating and cooling energy consumptions. The second cooling typing was used to set the default cooling equipment type for houses where the climate zone indicated presence of cooling. In one climate zone, ZONE 9, the analysis yeilded a most common configuration of heatpump for heating and cac for cooling. In this instance our judgement was that this didn't make sense for home modeling and might well be an error in the survey (possibly a misunderstanding on the part of the interview subject) and the cooling equipment was also set to HP.

RECS uses coding variables for age of homes. In order to get average ages of home within regions dates had to be assigned from the categorical data. The RECS variable YEARMADE was re-coded (as YRBUILT), as shown below. In each case (except BEFORE 1940) the mid-point of the category was used to represent the year. Since "before" has no mid-point that method was not applicable. 1925 was selected as a somewhat arbitrary value to represent this group.

Foundation types were coded based on several RECS variables. RECS categorizes home by several yes/no variables related to foundations: CRAWL,CELLAR, and CONCRETE. These were re-coded into a single variable. In about 700 cases where more than one type is reported the assignment is ordered to give precedence to basements—if anything else is reported along with a basement the home is coded as having a basement. RECS has several levels of basement heating but for this analysis any heated area in the basement was coded as "conditioned basement".

RECS categorizes homes as being 1, 2, 3, 4 OR MORE, SPLIT LEVEL, or OTHER. The first four of these were re-coded as 1,2, 3 and 4 respectively. There are actually very few single family homes of more than 4 stories so these were ignored. SPLIT LEVEL, were coded as two stories. It is unclear from the RECS documentation what OTHER really means—possibly geodesics or A-frames. These were coded as single floor, which is different from last update where they were dropped.

The "input" variables were the typical value for the climate zone. For numeric variables typical configuration was calculated as the mean value. For qualitative or Y/N variables we calculated the mode.


Consumptions were calculated in two phases. Space heating, space cooling and water heating consumption were calculated based on the most frequent type of Space heating fuel X Water heating fuel X Central Present (or not), within each climate zone. Appliance consumptions were the average for fuel and climate zone. We looked at calculating consumption separately for space heating and water heating, which would have had the advantage of basing the averages on larger numbers of homes, but some of the results were counter intuitive—the "typical" house had different fuels for space and water heating.

Space heating consumption is based on the RECS variables BTUELSPH, BTUNGSPH, BTUFOSPH, the fuel specific space heating energy; electric, natural gas, and fuel oil.

Water heating consumption is based on the RECS variables BTUELWTH, BTUNGWTH, BTUFOWTH, the fuel specific water heating energy; electric, natural gas, and fuel oil.

Electrical consumption for appliances is somewhat complex. The existing defaults analysis contained two variables related to electrical appliances: BUTELAPP and BTUELMIS, which are major appliance and misc appliances respectively. The RECS variable for appliances is BTUELAPL which is all electrical appliances except refrigerators. BTUELAPP is the sum of refrigerators, freezers, clothes washers, dishwashers, and electrical cooking. BTUELMIS is BTUELAPL less freezers, clothes washers, dishwashers, and electrical cooking.

RECS no longer reports electrical cooking separately, which posed a problem for calculation BTUELMIS. For this update electrical cooking energy was taken from RECS 1993. The RECS 1993 variable BTUELCOK was analyzed by the combination of Census Division + the four largest states (total of 13 regions), and housing type, where electricity was the main fuel used to cook. The resulting consumption figures were added to the RECS 2001 data set. Use of 1993 data is not ideal, but no better national level estimates were immediately available.

Natural Gas appliance are much more straightforward. This consumption is represented by the RECS variable, BTUNGAPL appliances; mostly water heaters and cooking, some clothes dryers) which is used without modification.

RECS has not included data on lighting consumption since 1993, so those values continued to be used.

For all RECS consumptions: where there is missing data or no answer RECS uses the coding 9999999 (or similar, the exact number of digits can vary). In this analysis missing values were set to 0. This allows averaging of minority fuels and prevents a small number of homes with high use of that kind (e.g. auxiliary elec. heating in a home with main fuel of NG) from skewing the average.


The savings methodology was not specifically updated in this revision. Savings are estimated by applying a percentage of improvement to the end-use consumptions space heating, space cooling, water heating, appliances, and lighting.


AHAM, "Energy Efficiency and Consumption Trends, var. products." July 1996

Birdsall, B., W.F. Buhl, K.L. Ellington, A.E. Erdem, and F.C. Winkelmann. 1990. "Overview of the DOE-2 building energy analysis program, version 2.1D." LBL-19735, Rev. 1. Berkeley, CA: Lawrence Berkeley National Laboratory.

Jennings, J., M. Moezzi, R. Brown, E. Mills, R. Sardinsky, B. Heckendorn, D. Lerman, L. Tribwell. 1997. "Residential Lighting: The Data to Date." Journal of the Illuminating Engineering Society. Vol. 26, No. 2. Summer 1997. also LBL-35484, March, 1996.

Lutz, J.D., Liu, X., McMahon, J.E., Dunham, C., & McGrue, Q.T. "Modeling Residential Hot Water Use Patterns" No. LBL-37805. Lawrence Berkeley National Laboratory. November 1996.

Mills, E.. et al. 2007. "Home Energy Saver: Documentation of Calculation Methodology, Input Data, and Infrastructure." Lawrence Berkeley National Laboratory Report No. 51938. [PDF]

Mills, E. 2004. "Inter-comparison of North American Web- and Disk-based Tools for Residential Energy Analysis." Energy and Buildings 36:865-880. [PDF] A more extensive version was previously published as "Review and Comparison of Web- and Disk-based Tools for Residential Energy Analysis." LBNL-50950. September 5, 2002. [PDF]

Sanchez, Marla C., "Miscellaneous Electricity Use in U.S. Residences." MS Thesis, University of California, Berkeley. June 1997.

Tribwell, Lyle S. and David I. Lerman. 1996. "Baseline Residential Lighting Energy Use Study." Proceedings of the 1996 ACEEE Summer Study on Energy Efficiency in Buildings. p. 3.153.

US DOE. 1996. "Electric Power Annual 1995." Energy Information Administration. DOE/EIA-0348(95)/1,2. July, December.

US DOE. 1994. "Emissions of Greenhouse Gases in the United States, 1987-1992." DOE/EIA-0573. October.

US EPA. 2009. "eGRID2007 Version 1.1" Washington, DC: U.S. Environmental Protection Agency.