Research On Dynamics Modeling And Analysis Of Occupant Energy Consumption Behavior In Office Buildings

Building occupant behavior is one of the main reasons for difference in building energy performance.How to quantify the uncertainty of building occupant behavior is important for energy performance simulation.Existing research has carried out a lot of work in the modeling and simulation of energy-related building occupant behavior,and most of them have established typical occupant behavior models for adaptive behavior of the built environment.However,the diversity and complexity of non-physical environmental factors increase the difficulty of occupant behavior modeling,and there are few studies focus on modeling autonomous driven occupant behavior.The generation mechanism behind autonomously driven occupant behavior is still in the exploratory stage.Therefore,this paper introduces the interdisciplinary research method of "system behavior" to explore the modeling method of autonomously driven occupant behavior,and further understand the generation mechanism of autonomous driven occupant behavior.This paper takes the autonomously driven occupant energy consumption behavior in a large government office building as the research object.First,the room-level active power data is used to indirectly characterize the energy consumption behavior of indoor occupants.Then,based on the theory of human dynamics,a dynamic modeling method for the occupant energy consumption behavior of office buildings was proposed.The overall occupant energy consumption behavior dynamic model of the sample offices and the respective occupant energy consumption behavior dynamic model of different offices are established,and the parameter laws and influencing factors of the occupant energy consumption behavior dynamic model of different offices are analyzed.The modeling results show that the time statistical characteristics of occupant energy consumption behavior of office buildings have significant non-Poisson characteristics.The overall occupant energy consumption behavior dynamics model of the sample office is most suitable for fitting with a lognormal distribution of μ =1.10 andσ =1.50.The dynamic model of occupant energy consumption behavior in different offices can be better fitted with three distribution types: truncated power-law distribution,stretched exponential distribution and log-normal distribution.Among them,the truncated power law distribution model has the best general applicability,and the power exponent of the model is mostly approximately 1.In terms of influencing factors of the models,the number of office energy consumption events and electricity consumption have no significant influence on the power exponents of the truncated power law distribution model and the stretched exponential distribution model,but have a positive correlation with the exponent λ.The expectation of the lognormal distribution model is negatively correlated with the number of office energy consumption events and electricity consumption.The variance will first decrease and then increase with the increase of the number of energy consumption events and electricity consumption.Based on the dynamic modeling of occupant energy consumption behavior,this paper analyzes the burst and memory of the time series of occupant energy consumption behavior,and uses the priority task selection model and the cascaded inhomogeneous Poisson process model study the generation mechanism of occupant energy consumption behavior.The research results show that occupant energy consumption behavior has typical strong burst and weak memory,which is the same as most human behaviors.The paroxysmal approaching 1,and memory approaching 0.Both of the two generation mechanism models can reasonably explain the reason why the time statistical characteristics of occupant energy consumption behavior show truncated power law distribution.The generation mechanism of occupant energy consumption behavior may be rational decision-making or the periodic rhythm of the system itself.Finally,the research results of this paper can be applied to building performance simulation,building system design and building energy-saving operation.