Window Opening Behavior And Modelling Of Xi’an Residential Buildings In Different Seasons

Windowing behavior is one of the most common human behaviors in architecture.The study found that human behavior has a significant impact on building energy consumption,and it is also a key factor causing the uncertainty of building energy consumption.Windowing ventilation can not only change the balance of temperature and humidity in the room to improve air quality,but also reduce the energy consumption of the building.However,the windowing behavior is influenced by many driving factors and has strong uncertainty,which brings difficulties to the dynamic simulation of building energy consumption.Finding out the driving factors of windowing behavior and establish an accurate windowing behavior model is helpful to improve the accuracy of building energy simulation.Taking four different houses in Xi’an area as test objects,the indoor and outdoor environmental parameters and window switch status of seven rooms were monitored for one year.The influence of environmental factors in different seasons on window-opening behavior was analyzed by means of multi-factor variance method to obtain window opening duration and window opening probability in different seasons.The neural network model of windowing behavior in different seasons was established and verified.The final conclusions are as follows:(1)The window opening frequency of the bedroom is significantly higher than that of the living room,and the bedroom basically keeps opening the windows at 8:00 in the morning every day in four seasons,and the time of opening the window reaches its peak between 8:00 and 10:00.(2)The window opening probability of 8:00 a.m.in spring reaches a peak of 45%;in summer,a double peak of 27.5% occurs between 7:30am and 9:00am and 18:00 PM.The probability of opening Windows at 7:30-9:30 every day in autumn reaches a peak of 50%.In winter,the probability of opening Windows between 8:00 am and 10:00 am reaches a peak of 43%.(3)The window opening periods for the four seasons of living rooms are obviously distributed.The window opening time of the No.3 living room is spring,summer,winter and autumn from high to low.No.4 sitting room is summer,autumn,spring,winter in turn by tall arrive low,and the bedroom compares sitting room crisscross fluctuation.(4)Through the model of opening Windows,it is found that the proportion of driving factors influencing the switched windows in different seasons is indeed different,and the driving forces of each influencing factor are variable with the change of seasons.(5)A total of 28 probability models of window opening in different seasons in seven rooms were established by neural network method,and the accuracy of the models was above 70%.The introduction of these models in the dynamic simulation of building energy consumption was helpful to improve the accuracy of the simulation.