A numerical simulation of measured transient temperatures in the walls, floor and surrounding soil of a buried structure

Abstract

Presents the results of a numerical simulation of measured heat transfer through a region surrounding a buried structure. The model applied in the study is a widely used whole building thermal simulation program of a type which predicts the thermal response of structures for building services requirements. A multi‐dimensional numerical conductive heat transfer module has been added to this program but this does not specifically address earth‐contact heat flows. This work attempts to assess the accuracy of the overall package when predicting earth‐coupled heat transfer. It is common practice in the field of building services not to use specific earth‐contact models and so it is important to assess the likely errors thus involved. The predictions of the finite‐volume model are compared with one year of data from a basement test facility. The results are analysed using the Differential Sensitivity Analysis method and an attempt is made to correlate predictive errors with periods of rainfall and snow coverage. It seems that a purely conductive model may be capable, given accurate input data, of satisfactorily predicting the transient temperature variations in the soil/concrete envelope surrounding this structure for the period of the year when no snow coverage is present. However, if one is to accurately model regions of earth‐contact (particularly at shallow depths) in a climate in which rainfall and snow are significant then these influences should be explicitly modelled.