Off-campus UNL users: To download campus access dissertations, please use the following link to log into our proxy server with your NU ID and password. When you are done browsing please remember to return to this page and log out.
Non-UNL users: Please talk to your librarian about requesting this dissertation through interlibrary loan.
Addressing uncertainties in residential energy performance benchmarking and projecting through data mining approach
The goal of this research is to develop a framework for improving the reliability of life cycle energy assessment of residential buildings. The proposed research method is primarily focused on developing improved building envelope performance benchmarking model and utilizing stochastic models to achieve improved life cycle operation energy consumption prediction. The proposed framework was applied to houses in a Midwest residential community, for which historical energy consumption records of the houses are available. The results of the models were validated through infrared thermal inspections. The results show the benchmarking model can generate comparable results to infrared thermal inspections. The stochastic life cycle energy consumption projection model can produce more practical results compared to that of traditional deterministic models and present more uncertainty information for decision-making. In the newly developed benchmarking model, the advantages of existing benchmarking approaches for addressing spatial uncertainties (i.e. multiple regression, principal component analysis and data envelopment analysis) are combined and lead to an improved approach for objectively evaluating the building envelope thermal performance for realizing the effective energy retrofitting. In the developed stochastic model for predicting life cycle energy consumptions of residential buildings, temporal variations of building energy consumption are considered by using Markov chain and neural network models. Major limitations of the developed models are identified as the follows: First, the impacts of spatial variability (uncertainty) of external "noise" factors that lead to the variation of the residential energy performance (REP) have not been fully considered. Second, occupants' behavior related temporal variability of the residential energy consumption factors is not accurately quantified in the model. These identified limitations are subject to further research contingent upon data availability.
Wang, Endong, "Addressing uncertainties in residential energy performance benchmarking and projecting through data mining approach" (2013). ETD collection for University of Nebraska - Lincoln. AAI3551352.