大型干坞边坡变形及其神经网络预测模型

Artificial Neural Network Forecast Model for Slope Deformation of Large-Scale Dry-Dock

分析了影响干坞边坡变形的因素,包括土体强度、无护坡时间、放坡坡率、分层开挖数、分层开挖深度、开挖步长、降水深度和坡顶荷载.在此基础上并结合典型实测数据,建立了干坞边坡变形的神经网络预测模型,预测结果与实测结果一致.此外,还提出了放坡开挖边坡变形的警戒值、边坡变形判断模式及相应的控制措施.

Factors influencing the deformation of a dry-dock slope were analyzed. They are soil mass strength, time of non-protection slope, gradient of slope, number of layout excavation, depth of layout excavation, excavation step, rainfall depth and load on slope top. Based on the above and the typical deformation data measured in-situ, an artificial neural network model of predicting the deformation of a dry-dock slope was proposed, and the prediction result is consistent with the measured in-situ result. In addition, the vigilance values for slope deformation for the expanding slope technology, the four judgement modes for slope deformation and their corresponding control measures were put forward.