Modeling Human Behavior in Building Performance Simulation: Gaussian Process and Monte Carlo Approach to the Energy Simulation of Residential Buildings

Occupant behavior has a large impact on residentialbuilding energy use; nevertheless, behavior models in buildingenergy simulation need scrutiny of investigation and use ofoversimplified schedules results in a great deal of uncertainty inbuilding energy prediction; the modeling of accurate occupantschedules requires a complex dataset with long-term observations,which is concealed during early stages of building projects. Thisstudy seeks to estimate behavior-related building operationschedules based on a minimum number of observations, so thatenergy simulation is effectively involved in the early stages ofbuilding projects. To this end, Gaussian process (GP) regressionis applied to modeling five major occupant activities (spaceoccupancy, activity level, hot water use, appliance use, andlighting control) of a single-family house in the United State.Monte Carlo simulation with sampling from GP-based occupantschedules demonstrates large variability of energy simulationresults according to different human behaviors