1) Discusses issues involved in baselining energy use at either the whole facility level or at the building level when monthly utility bills are available. Describes the variable-base degree day (VBDD) and the monthly mean temperature (MMT) regression modelling approaches for normalising energy use for changes in outdoor temperatures from the baseline year to subsequent years. Proposes statistical equations to determine prediction uncertainty for individual monthly and annual time intervals for change-point energy models whose residual variance is not constant. 2) Allies these concepts to monthly utility electricity and gas use data from eight army installations, evaluates the VBDD and the MMT approaches and discusses salient results of the analysis. Concludes by illustrating the use of the statistical method described in part 1.

Units: I-P

KEYWORDS: year 1997, Calculating, energy consumption, comparing, statistics, computer programs, buildings, USA