Energy Benchmarking in U.S. Office Buildings

Overview

     The benchmarking spreadsheets provided allow you to identify where your
specific office building ranks relative to others.  They calculate the energy use
intensity (EUI) of your building, provide the typical (median) EUI for office
buildings with the same characteristics as yours, and identify where your building's
performance ranks compared to others.  They go beyond the customary normalization
by floor area and account for performance differences due to variations in worker
density, the number of personal computers, operating hours, occupancy type, and
heating fuel types.  Beyond floor area, these characteristics were found to be the
most common and most important drivers of electric and non-electric energy use in
US office buildings.  Climate impacts on energy use were less significant, in part
because analyses were conducted within regional census divisions.

Background

     There is an important difference in this approach in comparison to other
benchmarking approaches.  Your building is compared to others that have the same
characteristics you provide as input.  Thus, you are not comparing your building,
which may have a high worker density (an important driver of energy use in 7 of 9
census divisions), to others with medium or low worker densities.  Other important
drivers of energy use are also accounted for.  Wide variances in these drivers
can strongly impact the energy use in office buildings.  By accounting for these,
we avoid comparing office buildings that have sound reasons for higher energy use
to those that do not.  Using this approach, we improve our ability to compare
buildings leading to better identification of problem buildings and more reliable
targets for our efficiency and cost reduction efforts.

Development Basis

     The EUI distributions used in these spreadsheets were developed based upon
statistical analysis of approximately 1500 office buildings in the US Energy
Information Administration's 1992 CBECS database.  These were divided into their
corresponding nine US census divisions for analysis.  Thus, different areas of the
US could have different results depending on what characteristics were found most
important to the locale.  A subset of over 70 building characteristics from the
CBECS database were selected and examined for their relationship to office building
energy use.  These were refined down to four characteristics that were the most
important determinants of electricity use and the four most important for 
non-electric energy use. These few characteristics explained most of the variations
in energy use that could be explained by considering all characteristics that had
statistically significant relationships to energy use.  Thus, addressing additional
characteristics provided limited value.  Within census divisions, climate was not
a major driver of either electric or non-electric energy use.

     Feedback on your experience using these spreadsheets is requested.  Please
direct comments to Terry Sharp, Oak Ridge National Laboratory, Building 3147, PO Box
6070, Oak Ridge, TN  37831-6070; e-mail: sharptr@ornl.gov; telephone: (865) 574-3559.