Using Regression Equations to Determine the Relative Importance of Inputs to Energy Simulation Tools

摘要

A set of statistical regression equationsrnwas developed to predict relative heatingrnand cooling loads of external zones ofrncommercial buildings. The equations werernderived from the coil loads predicted byrnseveral thousand DOE-2 simulations. Thesernequations formed the basis for thernbuilding envelope criteria in ASHRAE/IESrnStandard 90.1-1989, "Energy EfficientrnDesign of New Commercial Buildings ExceptrnLow-Rise Residential Buildings." Becausernthese equations predict relative loads,rnthey can be used to determine the relativernimportance of a broad range of envelopernparameters across a variety of climaterntypes.rnThis paper presents the procedure used torndevelop the equations. The relativernimportance of all the major loads inputrnvariables are discussed for a samplernoffice building, for a broad range ofrnclimates. A load sensitivity analysis isrnthen performed, which permits directrncomparison of key envelope parameters.rnThe analysis results provide generalrnguidance to DOE-2 users as to thernrelative importance of specific loadsrninput variables.