Simulation and case study on residential stochastic energy use behaviors based on human dynamics

Simulating the stochastic energy use behaviors of occupants plays an important role in the accurate prediction and operational optimization of building energy consumption. There are many types of energy use behaviors in residential buildings with strong randomicity and complex driving forces. A set of reasonable stochastic energy use behaviors simulation method with improved operability and prediction accuracy is required for the stochastic energy use behaviors. Therefore, a description method for residential energy use behaviors is proposed based on the theory of human dynamics. The residential energy use behaviors are divided into persistent behaviors and transient ones according to the method. The characteristics and regularity of these behaviors are described from the inherent decision-making mechanisms, which lays a theoretical foundation for modelling residential stochastic energy use behaviors in residential buildings. Consequently, a simulation model and algorithm for stochastic energy use behaviors are established based on subject and event mechanism. The model focuses on the entire generation process from occupants' movement to energy use behaviors. The occupants' movement sequences and persistent event sorting units are generated according to the BARABASI priority queue model and the highest priority decision. Also, based on the various control logic, the occurrence of transient behaviors is predicted. Finally, a household is selected for the annual measurement of stochastic energy use behaviors. Based on the above simulation method, MATLAB and NETLOGO are utilized with algorithm programming to realize the dynamic simulation of stochastic energy use behaviors in this household. The accuracy of the model is verified after comparing with the measured data. The research results suggest that the simulation method can accurately predict residential stochastic energy use behaviors and realize the simulation visualization. (c) 2020 Elsevier B.V. All rights reserved.