摘要
基坑工程危险性大、安全系数相对较低,为保障施工与后期建设安全,需要密切监测基坑支护桩位移量,并在监测基础上选择合理模型进行预测。本文针对灰色马尔可夫模型数据量要求小但精度不高,BP神经网络模型精度较高但对数据量要求大的特点,提出灰色马尔可夫-BP神经网络组合模型预测深基坑支护桩位移。结果表明,组合模型外符合精度达到0.35 mm,相对2种单一模型均有较大程度提高。
Foundation pit engineering is more dangerous and the safety factor is relatively low,in order to ensure the safety of construction and later construction,it is necessary to closely monitor the displacement of foun dation pit supporting piles and select a reasonable model for prediction on the basis of monitoring.In this paper,according to the characteristics of low precision required by Grey Markov model and high precision required by BP neural network model,the grey Markov BP neural network combined model is proposed to predict the displacement of supporting pile in deep foundation pit.The results show that the accuracy of the first mock exam is 0.35 mm and the two models are much better than others.
作者
俞校飞
俞明
YU Xiaofei;YU Ming(Zhejiang Huajia Technology Co.,Ltd.,Hangzhou 310051,China;Zhejiang Guokun Construction Group Co.,Ltd.,Hangzhou 311200,China)
出处
《测绘与空间地理信息》
2022年第8期207-209,219,共4页
Geomatics & Spatial Information Technology
关键词
灰色马尔可夫
BP神经网络
深基坑
支护桩位移
Grey Markov
BP neural network
deep foundation pit
displacement of supporting pile