When on-site tests are performed in artificial ground freezing projects, the conventional analytical method for determining the soil temperature field often presents significant error under certain conditions, particularly when the temperature-measuring pipe is far removed from the freezing front. This paper proposes a new method—called the undetermined coefficient method (UCM)—that more accurately predicts the actual soil temperature field and position of the freezing front. The UCM comprises a new mathematical formula developed based on a series of model tests performed under three different conditions—no seepage, constant seepage, and sudden seepage—and a multiple data regression analysis-based method for determining the coefficients in the developed formula. The proposed UCM was compared to a conventional analytical method based on the results from model tests and from the construction of the cross-channel of Guangzhou Metro Line 11. The results indicated that the UCM presented higher stability and practical value than conventional alternatives and could accurately predict the actual soil temperature field and position of the freezing front. Therefore, the UCM is expected to be of high practical value in artificial ground-freezing operations.

当在人工地面冻结项目中进行现场测试时，确定土壤温度场的常规分析方法在某些条件下通常会出现很大的误差，尤其是当温度测量管距离冻结前沿很远时。本文提出了一种称为不确定系数法（UCM）的新方法，该方法可以更准确地预测实际的土壤温度场和冻结前沿的位置。UCM包括一个新的数学公式，该公式基于在三种不同条件（无渗水，恒定渗水和突然渗水）下进行的一系列模型测试而开发，以及用于确定已开发公式中系数的基于多数据回归分析的方法。根据模型测试和广州地铁11号线交叉通道的建设结果，将拟议的UCM与常规分析方法进行了比较。结果表明，UCM具有比常规替代方案更高的稳定性和实用价值，并且可以准确地进行计算。预测实际的土壤温度场和冰冻锋线的位置。因此，预期UCM在人工地面冻结操作中具有很高的实用价值。