Foundation pit excavation will cause the surrounding soil to settle, and may even cause disasters such as cracks in adjacent buildings or underground pipelines, so it is of great significance to predict the deformation of foundation pit in advance for the safety of foundation pit engineering. In order to achieve high-precision prediction of foundation pit surface settlement, this paper utilized convolutional neural network (CNN), BP (back-propagation) neural network and long short term memory (LSTM) neural network algorithms to construct a foundation pit settlement deformation prediction model. Based on the entropy method and the CRITIC weight method, three machine learning algorithms were combined to establish a combinatorial prediction model. Based on the foundation pit deformation monitoring data, the accuracy of the single prediction model and the combined prediction model used in this paper was evaluated. The results show that compared with the prediction model of a single machine learning algorithm, the average absolute error of the combined entropy method prediction model is reduced by 90.79%, the mean square error by 99.44%, and the average absolute percentage error by 90.33%. The average absolute error of the combined prediction model of the CRITIC method was reduced by 86.40%, the mean square error was reduced by 98.31%, and the average absolute percentage error was reduced by 84.94%. The combined prediction model proposed in this paper has better prediction performance than that of the single model, and the research is of great significance to improve the prediction accuracy of foundation pit surface settlement and ensure the safety of foundation pit construction.
为克服单一预测模型在适应性、预测精度和稳定性等方面的缺陷,本文以沈阳市沈河区方家栏车站为研究对象,结合卷积神经网络(convolutional neural network,CNN)、BP(back-propagation)神经网络、长短期记忆(long short term memory,LSTM)神经网络三种算法模型,基于熵值法和CRITIC法建立组合预测模型,并对基坑施工引起的周边地表沉降进行预测分析。通过与实测沉降值对比,结合多个指标评估组合预测模型的精度和可靠性,为机器学习模型在基坑周围地表沉降方面的预测提供参考。
WANGP, MAS Y, YUEZ W,et al.A theoretical study on the spatial effect of water-rich foundation pit instability failure[J].AIP Advances, 2021,11(1): 015049.
[2]
YANGW L, HUY, HUC,et al.An agent-based simulation of deep foundation pit emergency evacuation modeling in the presence of collapse disaster[J].Symmetry,2018,10(11):581.
[3]
LIUL L, WUR G, CONGRESSS S C,et al.Design optimization of the soil nail wall-retaining pile-anchor cable supporting system in a large-scale deep foundation pit[J].Acta Geotechnica,2021,16(7):2251-2274.
HEZhiyong, ZHENGWei.Deformation prediction of deep foundation pit based on BP neural network[J].Journal of South China University of Technology (Natural Science Edition),2008,36(10):92-96.
LIUCong, HEYueguang, SHAOLeisen.Neural network prediction for horizontal displacement of deep foundation by MATLAB toolbox[J].Mining and Metallurgical Engineering,2016,36(5):27-29.
SONGChuping.Improved BP neural method for deformation predication of deep excavation[J].Journal of Civil Engineering and Management,2019,36(5):45-49, 55.
[10]
OUYANGX, NIEJ W, XIAOX.Study of improved grey BP (back propagation) neural network combination model for predicting deformation in foundation pits[J].Buildings,2023,13(7):1682.
WANGJunbao, LIUXinrong, LIPeng,et al.Study on prediction of surface subsidence in mined-out region with the MMF model[J].Journal of China Coal Society,2012,37(3):411-415.
TANPeng, CAOPing.Predicting surface settlement of tunnel using grey relational-support vector machine[J].Journal of Central South University (Science and Technology),2012,43(2):632-637.
[17]
LIH L, ZHAOZ Z, DUX.Research and application of deformation prediction model for deep foundation pit based on LSTM[J].Wireless Communications and Mobile Computing,2022,2022:9407999.
HONGYuchao, QIANJiangu, YEYuanxin,et al.Application of CNN-LSTM model based on spatiotemporal correlation characteristics in deformation prediction of foundation pit engineering[J].Chinese Journal of Geotechnical Engineering,2021,43():108-111.
[22]
YUANZ, GAOL, CHENH J,et al.Study on settlement of self-compacting solidified soil in foundation pit backfilling based on GA-BP neural network model[J].Buildings,2023,13(8):2014.
[23]
ZHANGQ, MAY N, ZHANGB,et al.Time series prediction on settlement of metro tunnels adjacent to deep foundation pit by clustering monitoring data[J].KSCE Journal of Civil Engineering, 2023,27(5):2180-2190.
[24]
HANY L, WANGY, LIUC Y,et al.Application of regularized ELM optimized by sine algorithm in prediction of ground settlement around foundation pit[J]. Environmental Earth Sciences,2022,81(16):413.
[25]
LIUY H, FENGX.Research on prediction of ground settlement of deep foundation pit based on improved PSO-BP neural network[C]//2021 5th International Conference on Water Conservancy,Hydropower and Building Engineering,May 28,2021,Qingdao,China,EDP Sciences,2021:1-4.
GEChangfeng, HUQingxing, LIFangming.A study on application of artificial neural network in prediction of ground surface settlement around deep foundation pit[J].Journal of Disaster Prevention and Mitigation Engineering,2008,28(4):519-523.
[28]
LECUNY, BOTTOUL, BENGIOY,et al.Gradient-based learning applied to document recognition[J].Proceedings of the IEEE,1998,86(11): 2278-2324.
[29]
RUMELHARTD E, DAVIDE, MCCLELLANDJ L,et al.Parallel distributed processing:explorations in the microstructure of cognition [M].Cambridge:MIT Press,1986:282-285.
SHERSTINSKYA.Fundamentals of recurrent neural network (RNN) and long short-term memory (LSTM) network[J].Physica D:Nonlinear Phenomena,2020,404:132306.
XUHousheng, GUOJiali.Short-term traffic flow forecast based on Bi-GRU-Atten algorithm with multi-layer time attention mechanism[J].Journal of Liaoning Technical University(Natural Science),2023,42(6):763-768.
CHENWanzhi, LIHaozhe, LIUHengjia,et al.Wind turbine fault prediction method based on LSTM and optimized SVM[J].Journal of Liaoning Technical University (Natural Science),2022,41(4):379-384.
LUZhifan, ZHAOQian.Short-term power load forecasting method based on ICEEMDAN-DCN-Transformer[J].Journal of Shenyang University of Technology,2024,46(4):388-396.