A case study on thermal conductivity characteristics and prediction of rock and soil mass at a proposed ground source heat pump (GSHP) site.

Shallow geothermal energy (SGE) has a wide range of applications in the field of building cooling and heating. Ground source heat pump (GSHP) system is a technology to extract SGE. The design of borehole heat exchanger (BHE) has a great impact on heat transfer performance and investment cost, so it is important to accurately measure the thermal conductivity of rock and soil. Therefore, this study conducted field in-situ thermal response test (TRT) and laboratory sample test based on distributed optical fiber temperature sensor (DOFTS) in LY research area of Changchun, Northeast China. After comparing the differences and analyzing the reasons, an in-situ thermal conductivity prediction model was established based on artificial neural network (ANN) algorithm to predict in-situ thermal conductivity based on basic physical property parameters of laboratory tests. This model is used to supplement the layered thermal conductivity lacking in the CY study area. The results show that the distributed thermal conductivity can be obtained and the layered thermal conductivity can be calculated by improved combined thermal response test (ICTRT). The average layer thermal conductivity of laboratory test is about 12.2% lower than that of field test, but the thermal conductivity of the two test methods has the same variation trend along the depth. The thermal conductivity of rock mass is positively correlated with water content, negatively correlated with porosity and positively correlated with density. The result error of the in-situ thermal conductivity prediction model established by calculation is mainly within ± 5%, which is reliable and accurate. This model is used to supplement the layered thermal conductivity of the CY01 test hole. The research results can provide a new way to determine the thermal conductivity in SGE exploration.