Using regression equations to determine the relative importance of inputs to energy simulation tools

A set of statistical regression equations was developed to predict relative heating and cooling loads of external zones of commercial buildings. The equations were derived from the coil loads predicted by several thousand DOE-2 simulations. These equations formed the basis for the building envelope criteria in ASHRAE/IES Standard 90.1-1989, "Energy Efficient Design of New Commercial Buildings Except Low-Rise Residential Buildings." Because these equations predict relative loads, they can be used to determine the relative importance of a broad range of envelope parameters across a variety of climate types. This paper presents the procedure used to develop the equations. The relative importance of all the major loads input variables are discussed for a sample office building, for a broad range of climates. A load sensitivity analysis is then performed, which permits direct comparison of key envelope parameters. The analysis results provide general guidance to DOE-2 users as to the relative importance of specific loads input variables.