Study On Impacts Of Indoor Office Environment On Occupants’ Sensation And Satisfaction

A healthy,comfortable,and productive indoor environment is an essential element of humans’well-being in a modern society.As an interdisciplinary field,indoor environment quality（IEQ）contains four factors–thermal environment,visual environment,acoustic environment,and indoor air quality（IAQ）,which have been studied separately in most of the previous studies.Thus,the interactions between environmental factors and the impact of overall indoor environment on occupants still need to be discussed.Meanwhile,with the rapid development of environmental monitoring technology,how the environmental information impact occupant’s perception is becoming a crucial issue.This dissertation focuses on the impact of IEQ,including thermal,acoustic,visual and IAQ,on occupants’sensation and satisfaction with indoor environment.The following contents were discussed.First,we investigated IEQ in five types of non-residential buildings,including office,school,hotel,shopping mall and hospital,in Chongqing,China.The investigation was conducted by measuring physical environmental parameters of buildings and surveying occupant’s sensation and satisfaction with indoor environment.The results indicated that qualification rate of thermal environment was higher in winter than in summer.The most thermal complaint was about non-uniform temperature distribution and dryness.About28%of measured area were found to have lower illuminance than requirement of Chinese standard,while the illuminance was significantly higher in shopping mall and that led to a significant complaint with glare.Excessive noise was found in all of the buildings and the sound pressure was higher in summer than in transition seasons.The classroom,underground supermarket,and hospital ward were found to have excessive concentration of CO₂.All of the investigated buildings failed to effectively filter outdoor PM_2.5 and that caused a very significant correlation between indoor and outdoor PM_2.5 concentration.Acoustic environment was found to have significant lower satisfaction than the other three environment factors.Second,we exposed subjects to a simulated office environment in a IEQ lab and studied their perception of four IEQ factors.The 31 simulated IEQ conditions were designed by orthogonal design and uniform design.The partial correlation analysis and regression analysis were applied to identify interactions between the four IEQ factors.The results indicated that illuminance had a significant impact on acoustic perception especially when sound pressure was low.The acoustic sensation was quieter and acoustic satisfaction was higher when illuminance was low.The results also verified the correlation between PMV and IAQ perception that the IAQ perception significantly decreased with PMV.Meanwhile,we found a slight negative impact of sound pressure on IAQ perception.To study the impact of environmental information on occupants’perception,we exposed subjects to several IEQ conditions and informed them different environmental information,including quantitative information of parameters,qualitative information of environment ratings and information better than actual conditions.Paired t test was applied to identify whether subjects changed their satisfaction and sensation when they were informed of different information.The results indicated that quantitative information only influenced IAQ perception but not thermal,acoustic,and visual.The qualitative information had a wider impact on perception of four environment factors,positively and negatively.The better environment information significantly improved subjects sensation and satisfaction with IEQ under most of the simulated conditions.At last,we interpreted the relationship between overall satisfaction and thermal,acoustic,visual and IAQ satisfaction by building three statistical model and two machine learning model.We evaluated the performance of resulting models in terms of prediction accuracy,stability,and interpretability.The SHAP analysis was applied to quantify contribution of four single-domain satisfactions to overall satisfaction.The results indicated that the weighted harm mean model was the best model as it had an accuracy close to machine learning mode,and better stability and interpretability.The model revealed a non-linear correlation between overall satisfaction and single-domain satisfaction that a single-domain satisfaction was more influential when it was low than when it was high.