Evaluation of Bias Issues within Regression-Based Inverse Modeling Methods Against Climate and Building Characteristics Using Synthetic Data
Buildings; Energy Efficiency
Typically, a model of the energy use of a building is created using the building's characteristics and climate information for its location. However, models of existing homes can be created using monitored energy and climactic data in a process called inverse modeling. Such a technique has potential for those seeking to evaluate the savings a home before and after a retrofit (Meier, Busch, & Conner, 1988).Several programs use an inverse modeling technique to model a home's energy use based on total home energy use, heating and cooling energy use, and indoor/outdoor temperature data from a home (Kissock, Haberl, & Claridge, 2003). For example PRISM, the PRInceton Scorekeeping Method, was used widely in the 1980s and 90s to evaluate the performance of residential energy efficiency improvements (Fels, 1986). In 2006, researchers from Building America evaluated several homes using a least-squares regression technique on monitored data from these homes. The researchers compared the performance of low-energy homes working towards the Building America goal of 70% whole-house efficiency to homes built to minimum code requirements. The 2006 study found a correlation between the efficiency of a home to the performance reported by the least-squares regression, but asserted that more analysis into the method was needed in order to ascertain its accuracy, especially regarding floor type, climate, and house size (Chasar, et al., 2006).This paper seeks to evaluate the accuracy of the regression technique used in 2006 by using the same least-squares regression analysis on data created from hourly energy simulation software. Synthetic data allows researchers to cheaply and quickly obtain more data for analysis and to isolate the effects of single characteristics on a home's performance in a regression analysis model.
Buildings - Energy Efficiency
Florida Solar Energy Center and Cummings, Jamie, "Evaluation of Bias Issues within Regression-Based Inverse Modeling Methods Against Climate and Building Characteristics Using Synthetic Data" (2010). FSEC Energy Research Center®. 304.