부산대학교 도서관

For measurement and verification (M&V) of energy savings in buildings, we propose an approach based on Gaussian Process (GP) modeling that can represent nonlinear energy behavior, multivariable interactions, and time correlations while quantifying uncertainty associated with predictions. We applied GP modeling to determine energy savings from BuildingIQ's energy management system deployed at the Advanced Photon Source Office building at Argonne. The case study demonstrates the potential strengths of GP models for M&V and explores the importance of dataset characteristics and explanatory variables for the reliability of analysis results. The case study illustrates the capability of GP modeling to predict hourly dynamic behavior, exploiting the possibility to reduce uncertainty in energy-use predictions using measured data with finer time resolutions. The proposed M&V approach is amendable to automation in energy management systems and continuous monitoring of energy performance. [ABSTRACT FROM AUTHOR]