MULTIVARIATE REGRESSION MODELING

An empirical or regression modeling approach is simple to develop and easy to use compared to detailed hourly simulations of energy use in c ommercial buildings. Therefore, regression models developed from measu red energy data are becoming an increasingly popular method for determ ining retrofit savings or identifying operational and maintenance (O&M ) problems. Because energy consumption in large commercial buildings i s a complex function of climatic conditions, building characteristics, building usage, system characteristics and type of heating, ventilati on and air conditioning (HVAC) equipment used, a multiple linear regre ssion (MLR) model provides better accuracy than a single-variable mode l for modeling energy consumption. Also, when hourly monitored data ar e available, an issue which arises is what time resolution to adopt fo r regression models to be most accurate. This paper addresses both the se topics. This paper reviews the literature on MLR models of building energy use, describes the methodology to develop MLR models, and high lights the usefulness of MLR models as baseline models and in detectin g deviations in energy consumption resulting from major operational ch anges. The paper first develops the functional basis of cooling energy use for two commonly used HVAC systems: dual-duct constant volume (DD CV) and dual-drier variable air volume (DDVAV). Using these functional forms, the cooling energy consumption in five large commercial buildi ngs located in central Texas were modeled at monthly, daily, hourly, a nd hour-of-day (HOD) time scales. Compared to the single-variable mode l (two-parameter model with outdoor dry-bulb as the only variable), ML R models showed a decrease in coefficient of variation (CV) between 10 percent to 60 percent, with an average decrease of about 33 percent, thus clearly indicating the superiority of MLR models. Although the mo dels at the monthly time scale had higher coefficient of determination (R-2) and lower CV than daily, hourly, and HOD models, the daily and HOD models proved more accurate at predicting cooling energy use.