基于 BP 人工神经网络的 深基坑围护结构水平位移预测研究

在深基坑施工过程中，对围护结构进行水平位移监测是保证施工安全的重要措施之一，而分析预测围护 结构的变形趋势更是重中之重。为此，对围护结构自动化监测设备进行实地调研，提出结合 BP 人工神经网络模 型对围护结构水平位移进行多步滚动预测的方法。以南宁地铁 5 号线车站深基坑施工围护结构的真实监测数据为 训练样本，对样本数据分别进行 3 种模式学习：第一种，不同桩学习后，对在同一时间的预测结果作对比；第二 种，同一根桩在不同时间的间隔样本学习后，对在同一时间的预测结果作对比；第三种，同一根桩在实现多步滚 动预测后，对预测结果作对比。结果表明：3 种模式的预测误差均可满足要求，为实现围护结构变形自动预测提 供实用性强、可信度高的方法。

Automatic Prediction of the Deformation of a Retaining Structure in a Deep Foundation Pit Construction Based on a BP Artificial Neural Network

In the process of a deep foundation pit construction, horizontal displacement monitoring of retaining structures is an important measure to ensure construction safety, and the analysis and prediction of the deformation trend of the retaining structures are the most important. In this study, we conducted a field investigation on the automatic monitoring equipment of the enclosure structure and proposed a method for a multi-step rolling prediction of the horizontal displacement of the enclosure structure combined with the BP artificial neural network model. Real monitoring data of the deep foundation pit construction enclosure structure of the Nanning Metro Line 5 station were used as the training sample, and three modes of learning were performed on the sample data. The first was to compare the prediction results of different piles at the same time after learning; second, to compare the prediction results of the same pile at the same time after sample learning at different time intervals; and third, to compare the prediction results of the same pile after the multi-step rolling prediction. The results show that the prediction errors of the three modes satisfy these requirements. It provides a method with strong practicability and high credibility to realize the automatic prediction of the deformation of the enclosure structure.