The correlation between soil thermal resistivity and its main influencing factors was analyzed by reviewing the literature. To obtain an accurate prediction model of soil and carbonaceous rock thermal resistivity, the neural network feedback analysis technology was utilized and a prediction model developed. The laboratory results verify the effectiveness and superiority of the model. Dry density, saturation, and quartz were selected for the prediction model, which can comprehensively and reasonably reflect the main factors affecting soil and carbonaceous rock thermal conduction. Based on the comparison results between predicted and measured thermal resistivity, the proposed mode exhibited a satisfactory accuracy. Compared with the selected empirical relationships, the model had significant advantages in the prediction results for different types of soils.

通过查阅文献，分析了土壤热阻与其主要影响因素的相关性。为了获得准确的土壤和碳质岩热电阻率预测模型，利用神经网络反馈分析技术开发了预测模型。实验室结果验证了模型的有效性和优越性。预测模型选用干密度、饱和度和石英，能够全面合理地反映影响土壤和碳质岩热传导的主要因素。基于预测和实测热阻的比较结果，所提出的模式表现出令人满意的精度。与选定的经验关系相比，