A quantitative Modeling Method of Occupant Behavior of Office Building Based on Human Dynamics

Abstract

Occupant behavior has been proved to be the fundamental reason affecting the energy efficiency of buildings. Most of the previous researches on stochastic probabilistic modeling of occupant behavior usually simplify a single specific human behavior activity pattern to a completely random point process, and the modeling methods do not consider the coupling mechanism of behavioral events. In this paper, based on the theory of human dynamics, a modeling method of occupant behavior was proposed to describe the occupancy behavior pattern in the real situation of complex systems. In this case, the energy-using behavior events of electrical equipments were separated from the active power data of multiple single-person/multi-person offices. The time interval distribution models of the energy-using behavior of all equipment in all offices and the energy-using behavior of different office printers were established. It was found that the time interval distribution of all office equipment occupancy followed the lognormal distribution model, and the model index was independent of the time span. However, the time interval distribution of A4 laser printer usage behavior follows the stretched exponential distribution model, and the behavior in different offices is uncertain and does not comply with the traditional Poisson process hypothesis. The results show that there is an obvious generation rule of occupant behavior, and also a specific generation mechanism behind it. Finally, this paper discussed the application of a single specific time interval model of energyusing behavior and the extension of modeling method, which provides theoretical basis for occupant behavior prediction and building energy consumption simulation.