基于CGA模型的盾构扭矩预测研究  

Study on Shield Torque Prediction Based on CGA Model

作　　者：刘映晶 卢敬科[1] 陈城 刘维 LIU Yingjing;LU Jingke;CHEN Cheng;LIU Wei(Zhongtian Construction Group Co.,Ltd.,Hangzhou,Zhejiang 322199,China;School of Rail Transportation,Soochow University,Suzhou,Jiangsu 215000,China)

机构地区：[1]中天建设集团有限公司,浙江杭州322199 [2]苏州大学轨道交通学院,江苏苏州215000

出　　处：《河北工程大学学报（自然科学版）》2024年第2期51-58,共8页Journal of Hebei University of Engineering:Natural Science Edition

基　　金：国家自然科学基金资助项目(51978430);中天控股集团技术研发项目(ZTCG-GDJTYJS-JSKF-2021001)。

摘　　要：以盾构近距离下穿既有车站结构为背景,提出了一种结合卷积神经网络(Convolutional Neural Networks, CNN)、门控制循环单元神经网络(Gated Recurrent Unit, GRU)和注意力机制(Attention)的新型盾构荷载预测模型。首先用CNN-Attention模型提取数据的高维空间特征并区分不同特征的重要性,然后通过GRU模型提取数据的时序特性,紧接着通过注意力机制提取出重要时间节点信息,最后得出预测的结果。为验证所提模型的预测效果,选取了4种现有的算法进行比较。结果表明所提出的模型在三种评价指标上均优于其他算法模型,同时该模型还可为盾构刀具磨损、地表及结构变形等方面的预测研究提供思路。Shield load is an important parameter of shield machine,and accurate prediction of shield load is very important to ensure the safe construction of shield tunnel.In this paper,a new load predic-tion model(CGA),combining convolutional neural network(CNN),gate recurrent unit neural network(GRU)and attention mechanism(Attention),is proposed based on the shield machine cross existing station at close range.The CNN-Attention model is first used to extract the high-dimensional spatial fea-tures of the data and distinguish the importance of different features.Then the GRU model is used to ex-tract the temporal characteristics of the data,followed by the attention mechanism to extract the impor-tant time node information.Finally,the prediction results are obtained.To verify the prediction per-formance of the proposed model,four existing algorithms are selected for comparison.The results show that the proposed model in this paper outperforms other models in three evaluation metrics,and the pro-posed model can also provide reference for predicting researches on shield tunneling tool wear,surface and structural deformation,etc.

关 键 词：盾构隧道 扭矩预测 深度学习 注意力机制 时空特征 

分 类 号：TU472[建筑科学—结构工程]