机场高填方蠕变沉降的数值反演预测

将考虑时间效应UH模型编写用户材料子程序（UMAT）并嵌入到有限元软件ABAQUS中，再结合多岛遗传算法搭建了数值反演平台。在实际工程中降雨、融雪和临时堆载等因素都会扰乱常态蠕变的沉降规律，使得沉降加快形成突变沉降，进而影响数值反演的预测结果。通过引入非常态蠕变概念并分析突变期间沉降加速的机制，提出了考虑非常态蠕变的反演预测方法。此外，当填方体缺乏分区依据且以多个监测点的沉降数据作为反演目标时，引入了反演参数的分布函数来代替几何模型的分区。并在此基础上进一步考虑沉降突变期间的非常态蠕变影响，进而提出多监测点下考虑非常态蠕变的沉降预测方法。最后，通过河北省承德机场高填方含有沉降突变的监测数据，验证了多监测点数据下考虑非常态蠕变的数值反演的准确性和实用性。

Numerical back-analysis of creep settlement of airport high fill

In this paper, a user material subroutine (UMAT) of time-dependent UH model is written and embedded into the finite element software ABAQUS. Then a numerical back-analysis platform is established by combining multi-island genetic algorithm. In actual projects, factors such as rainfall, snowmelt and temporary loading will disturb the settlement law of normal creep. During this period of disturbance, the settlement rate is accelerated to form abrupt settlement, which has a great impact on the accuracy of numerical calculation. By introducing the concept of abnormal creep and analyzing the mechanism of settlement acceleration during this period, a numerical prediction method considering abnormal creep is proposed. When the fill body is lack of zoning basis and the settlement data of multiple monitoring points is taken as the back-analysis target, the distribution function of back-analysis parameters is introduced to replace the zoning of geometric model. On this basis, the influence of abnormal creep during the period of disturbance is further considered, and a numerical prediction method of settlement considering abnormal creep at multiple monitoring points is proposed. Finally, the accuracy and practicability of the numerical prediction method considering the abnormal creep under the condition of multiple monitoring points are verified by the settlement data of Chengde airport high fill.