Abstract Artificial ground freezing (AGF) is a method of paramount importance for underground construction in soft soil area. There were numerous studies concerned on field observations and the theoretical formulations on frost heaves. However, the field observations and validity of the theoretical models from practical points were lack. This paper focused on a more accurate practical prediction method of frost heave by multilayer field experiments and segregation potential (SP) model, for a strict deformation requirement AGF project in a non-stopping airport during construction. Thus, a large-scale area (3.5 times than actual frozen curtain) field experiment of multilayered temperatures and displacements was conceived and developed, to evaluate the freezing effect and deformation characteristics. Temperature variations show that the additional freezing tubes at the tunnel opening can only achieve faster cooling rate initially but could not descend the final stable temperature, and so substantially increasing freezing tubes at the tunnel opening may not be best choice to keep excavation safe. Multilayered temperatures and displacements show that deformation of multi-layers is directly relevant to the position of frost front. A simplified frost heave prediction method using segregation potential concept is proposed based on field experimental data. The temperature gradient and frost front function have been calculated based on field monitoring results and served for the model computation. This prediction method was further validated by both other’s published and our field experimental data. 2–8% relative errors surrounding the axis of frozen curtain prove the applicability of this practical prediction model is much well. This paper provides valuable reference for urban tunneling under extreme conditions such as non-stopping airport or adjacent to existing structures.

中文翻译：

---

人工冻结软土冻胀现场实测实用预测方法

摘要 人工地面冻结（AGF）是软土地区地下施工的一种重要方法。有大量关于现场观测和冻胀理论公式的研究。然而，缺乏从实践角度对理论模型的实地观察和有效性。本文针对建设期间对不经停机场的变形要求严格的 AGF 项目，通过多层现场试验和离析潜力 (SP) 模型，提出了一种更准确实用的冻胀预测方法。因此，构想并开发了大面积区域（实际冻结幕的3.5倍）多层温度和位移的现场试验，以评估冻结效果和变形特性。温度变化表明，在隧道洞口处增加冷冻管只能达到较快的初始冷却速度，而不能降低最终的稳定温度，因此大幅增加隧道洞口处的冷冻管可能不是保证开挖安全的最佳选择。多层温度和位移表明多层变形与霜锋位置直接相关。基于现场实验数据，提出了一种基于离析潜力概念的简化冻胀预测方法。根据现场监测结果计算了温度梯度和霜锋函数，并用于模型计算。这种预测方法得到了其他人发表的和我们的现场实验数据的进一步验证。围绕冻结帘轴的 2-8% 的相对误差证明该实用预测模型的适用性非常好。本文为机场直通或邻近既有建筑物等极端条件下的城市隧道掘进提供了有价值的参考。