Using multiple regression analysis to develop energy consumption indicators for commercial buildings in the U.S.

Multiple regression analysis plays an important role in evaluating the energy consumption of buildings. These models are commonly used to assess the energy performance of commercial buildings and to predict any potential for energy consumption reduction. In this study, the building simulation software DOE-2 was used to predict energy consumption. A total of 17 key building design variables were identified related to building envelope, building orientation, and occupant schedule. Since, building energy consumption depends on many operational and design parameters; large numbers of simulations are needed to generate data for the multiple regression models. To tackle this problem, a randomized approach was adopted to reduce the required number of simulations examining the whole design space. Monte Carlo simulation technique was used to generate thirty thousand combinations of design parameters, covering the full range for each climate region. In order to implement the Monte Carlo simulation, an in-house computer program was developed to interface with DOE-2 energy simulation software. Stepwise regression was used to reduce the number of parameters and only include the most effective parameters. R statistical analysis program was also used to develop the set of linear regression equations. Parametric study and sensitivity analysis between levels of most effective parameters were performed. The developing models can be used to estimate the energy consumption of office buildings in early stages of design.