Assessment of the U.S. Department of Energy's Home Energy Scoring Tool

Abstract

NOTICEThis report was prepared as an account of work sponsored by an agency of the United States government. Neither the United States government nor any agency thereof, nor any of their employees, makes any warranty, express or implied, or assumes any legal liability or responsibility for the accuracy, completeness, or usefulness of any information, apparatus, product, or process disclosed, or represents that its use would not infringe privately owned rights. Reference herein to any specific commercial product, process, or service by trade name, trademark, manufacturer, or otherwise does not necessarily constitute or imply its endorsement, recommendation, or favoring by the United States government or any agency thereof. The views and opinions of authors expressed herein do not necessarily state or reflect those of the United States government or any agency thereof. 1 In acknowledging individuals and organizations we do not mean to imply their endorsement of the research results. Our intention is simply to acknowledge their contributions and thank them. iv Executive SummaryThe National Renewable Energy Laboratory (NREL) conducted a series of assessments of the U.S. Department of Energy's (DOE) proposed Home Energy Scoring Tool (HEST). The primary objective of this work was to assess the accuracy of HEST as it was being developed and to provide information useful to DOE program managers and HEST development team at Lawrence Berkeley National Laboratory.NREL assessed the accuracy of HEST from the version used for the Home Energy Score pilot, released January 26, 2011, through the April 27, 2012 release. With the exception of Appendix A, Historical Progression of HEST Accuracy, this report reflects assessment of the April 27, 2012 release of HEST. Comparison of Predicted Energy Uses to Measured Energy UsesPredictions of electricity and natural gas (NG) consumption were compared with weathernormalized utility billing data for a mixture of newer and older homes located in Oregon, Wisconsin, Minnesota, North Carolina, and Texas. 2 The 859 electricity comparisons and 500 NG comparisons yielded the following:• HEST underpredicted electricity use by a median of 1%.• HEST underpredicted NG use by a median of 10%.The primary objective of the Home Energy Score program is to issue a score to the homeowner. The Score ranges from 1 to 10, where a home scoring a 1 uses the most energy and a home scoring a 10 uses the least. For 52% of the homes in this sample, the predicted Home Energy Score is within ±1 point of a score calculated from measured energy use. 3 Comparison of Predicted Energy Uses to Predictions From Other ToolsSimilar comparisons were made between predictions from two other commonly used residential energy analysis software tools, REM/Rate and SIMPLE, and weather-normalized utility billing data for the same set of homes. The results of the comparisons are presented along with those from HEST in Table ES-1 and Table ES-2. HEST energy use predictions compare well with the other two energy analysis software tools....

Field Assessment of Energy Audit Tools for Retrofit Programs

“…Changes to the Home Energy Score methodology since the pilot version have adjusted the model's energy calculations. For results using more recent versions of the Home Energy Score see Roberts et al 2012. For documentation of the changes made to the Home Energy Score see or the Home Energy Score documentation (LBNL 2010).…”

Scalable tuning of building models to hourly data

“…A 2008 pilot study [4] found 190 Home Energy Saver, REM/Rate, and SIMPLE residen-tial simulation models had a 25.1-96.6% error rate compared with actual monthly electrical energy usage. A 2012 study [5] found that 859 residential models across Home Energy Saver, REM/Rate, and SIMPLE had a mean absolute percentage difference of 24% compared with actual monthly electrical energy use and of 24-37% compared with actual natural gas use for a sample size of 500 houses. All of these studies use comparisons with monthly utility bill data; the challenge of accurately matching hourly or 15-minute data for dozens of sub-metered data channels is significantly more difficult.…”

“…Their usefulness is dramatically greater for existing buildings, for which existing data can be used to calibrate the energy model. However, differences between models and actual monthly utility bills on the order of 24-97% [4,5] are common. Many measurement and verification (M&V) protocols specify a required accuracy for a model to be legally useful; almost all organizations use ASHRAE Guideline 14 (subsection 5.3.2.4.f) which specifies a coefficient of variance for root mean squared error, RMSE, of <15% or 30% and a normalized mean bias error, NMBE, of <5% or 10% for calibrating to monthly or hourly data, respectively [6].…”

Field Assessment of Energy Audit Tools for Retrofit Programs

“…For results using more recent versions of the Home Energy Score see Roberts et al 2012. For documentation of the changes made to the Home Energy Score see or the Home Energy Score documentation (LBNL 2010).…”