| To reduce building energy consumption while ensuring personalized thermal comfort,creating a non-uniform indoor thermal environment is regarded as an effective method.Obtaining detailed indoor temperature distribution is the necessary basis for the design,evaluation and regulation of non-uniform indoor thermal environment.Computational fluid dynamics(CFD),as the most commonly used tool for indoor temperature distribution calculation,has high requirements for computing resources and computing time.Although many methods have been proposed to quickly predict the indoor temperature distribution,they are still not suitable for the real-time calculation in practical applications due to the difficulty of determining the heat sources.A method based on the Contribution Ratio of Indoor Climate(CRI)and finite air temperatures collected by fixed sensors has been proposed,which realizes the rapid prediction of indoor temperature distribution without determining the indoor heat sources in advance.However,its theoretical basis,prediction accuracy and application range need to be further discussed and improved.Therefore,this study proposes to combine CRI,mobile sensors and Proper Orthogonal Decomposition(POD)to achieve the dynamic prediction of indoor temperature distribution.Firstly,this study has provided a complete and detailed description of CRI,including a review of its existing contents,such as basic premises,definitions,calculation methods,etc.,and discussed its mathematical meaning for the first time.And two simple examples were given to illustrate how to calculate and understand CRI under forced convection and natural convection.On this basis,the form and number of heat sources in the existing algorithm were optimized.And a mobile sensor was used to collect the air temperature instead of the fixed sensor,combined with CRI to predict the indoor temperature distribution.Based on a typical office model,the effectiveness of using mobile sensors was discussed,and the influence of its acquisition height and acquisition distance on the prediction accuracy was analyzed as well.To break the limitation of CRI fixed value hypothesis on the application of this method,so as to further improve the accuracy and applicability,POD method which can effectively construct the mapping relationship between design parameters and design objectives was introduced to quickly obtain the dynamic distribution of CRI under any air supply conditions.In a simple office model with identified heat source conditions,the reliability of constructing CRI distribution by interpolation POD method was discussed,and the dynamic prediction of temperature distribution was realized combined with mobile sensors.The results showed that CRI is an effective index to describe the independent contribution of various heat sources to indoor temperature distribution.The improved algorithm is effective,and due to the limitations in practical application,using mobile sensors instead of fixed sensors can promote both reliability and accuracy on the basis of reducing the number of sensors.In the human activity area,the acquisition height of the mobile sensor has little effect on the prediction accuracy,while the acquisition distance should be big enough to make the distribution of acquisition points more dispersed.Besides,due to the influence of airflow distribution,the aquisition point should be far away from the area with complex airflow distribution,such as the room boundary.The introduction of POD method can break the application limitations caused by CRI fixed value hypothesis,realize the dynamic calculation of CRI distribution and improve the prediction accuracy of temperature distribution.However,in the area with complex airflow distribution,the calculation accuracy still needs to be improved. |
