基于NGO优化的CNN-BiLSTM-AM滑坡位移预测模型

王惠琴; 郭瑞丽; 何永强; 刘宾灿

doi:10.3969/j.issn.0258-2724.20240550

基于NGO优化的CNN-BiLSTM-AM滑坡位移预测模型

doi: 10.3969/j.issn.0258-2724.20240550

1.
兰州理工大学计算机与通信学院，甘肃兰州 730050
2.
西北民族大学土木工程学院，甘肃兰州 730030
3.
陕西建工安装集团有限公司，陕西西安 710068

基金项目: 国家自然科学基金项目（62261033，61861026）；甘肃省重点研发计划（21YF1GA381）；陕西省重点研发计划（2024GX-YBXM-42）

详细信息

作者简介:
王惠琴（1971—），女，教授，博士，研究方向为无线光通信理论与技术，智能信息与多媒体信号处理，E-mail：whq1222@lut.edu.cn

中图分类号: P642.22
计量
- 文章访问数: 42
- HTML全文浏览量: 36
- PDF下载量: 7
- 被引次数: 0
出版历程
- 收稿日期: 2024-10-26
- 修回日期: 2025-01-18
- 网络出版日期: 2025-07-16

NGO-Based CNN-BiLSTM-AM Model for Landslide Displacement Prediction

1.
School of Computer and Communication, Lanzhou University of Technology, Lanzhou 730050, China
2.
School of Civil Engineering, Northwest Minzu University, Lanzhou 730030, China
3.
SCEGC Installation Group Company LTD, Xi’an 710068, China

摘要

摘要:
针对目前滑坡位移预测研究中因采用单一预测模型而难以有效提取复杂序列特征，以及手动调整模型参数容易陷入局部最优等问题，提出一种基于北方苍鹰优化算法（NGO）的卷积-双向长短时记忆神经网络-注意力机制（CNN-BiLSTM-AM）滑坡位移预测模型. 首先，依据滑坡的影响因素，采用多元经验模态分解（MEMD）算法将多种滑坡位移数据分解为趋势项和周期项，其中，对趋势项位移，采用差分自回归移动平均（ARIMA）方法进行预测，对周期项位移，通过灰色关联度确定影响因素后，并构建CNN-BiLSTM-AM组合模型进行预测，其最优超参数通过NGO优化获得；其次，考虑周期项的滞后性，采用斯皮尔曼相关系数选取最优滞后期位移，进一步提升模型的预测效果；最后，利用甘肃渭源脱甲山滑坡监测数据进行验证. 结果表明：脱甲山滑坡总位移预测的均方根误差（RMSE）和平均绝对误差（MAE）分别低至0.22 mm和0.37 mm，展现出较高的预测精度；拟合优度（R²）达到0.98，充分验证所提模型在滑坡位移预测方面的有效性和可靠性.
- 滑坡位移预测 /
- 卷积神经网络 /
- 双向长短时记忆神经网络 /
- 北方苍鹰算法 /
- 注意力机制
Abstract:
A convolutional-bidirectional long short-term memory neural network-attention mechanism (CNN-BiLSTM-AM) prediction model optimized by the northern goshawk optimization (NGO) algorithm for landslide displacement was proposed to address challenges that a single prediction model fails to effectively extract complex sequence features and that manual parameter tuning tends to fall into local optima in current landslide displacement prediction research. Firstly, according to the factors affecting the landslide, the multivariate empirical mode decomposition (MEMD) algorithm was used to decompose various landslide displacement data into trend and periodic components. The trend components were predicted using the autoregressive integrated moving average (ARIMA) method. For the periodic components, influencing factors were identified through the gray correlation degree, and a CNN-BiLSTM-AM combined model was constructed for prediction. The optimal hyperparameters of this model were obtained through NGO. Then, by considering the lag of the periodic components, the Spearman correlation coefficient was used to select the optimal lagged displacement to further enhance the model’s predictive performance. Finally, the model was validated using monitoring data of the Tuojiashan Landslide in Weiyuan, Gansu Province. The results show that the RMSE and MAE of the total displacement prediction of the Tuojiashan landslide are as low as 0.22 mm and 0.37 mm, respectively, showing the prediction accuracy of the correction, while the R² reaches 0.98, which fully verifies the validity and reliability of the proposed model in landslide displacement prediction.
- landslide displacement prediction /
- convolutional neural network /
- bidirectional long short-term memory neural network /
- northern goshawk optimization algorithm /
- attention mechanism

HTML全文

图 1 滑坡位移预测流程

Figure 1. Flow chart of landslide displacement prediction

下载: 全尺寸图片幻灯片

图 2 CNN-BiLSTM-AM结构

Figure 2. CNN-BiLSTM-AM structure

下载: 全尺寸图片幻灯片

图 3 脱甲山滑坡监测点平面布置

Figure 3. Layout of monitoring points of Tuojiashan Landslide

下载: 全尺寸图片幻灯片

图 4 GP01监测点位移-降雨量监测曲线

Figure 4. Displacement–rainfall monitoring curves of GP01 monitoring point

下载: 全尺寸图片幻灯片

图 5 滑坡总体位移、垂直、水平位移MEMD分解结果

Figure 5. MEMD results of total, vertical, and horizontal landslide displacements

下载: 全尺寸图片幻灯片

图 6 总体趋势项位移预测结果

Figure 6. Overall displacement prediction results for trend term

下载: 全尺寸图片幻灯片

图 7 降雨-总体周期项位移时滞spearman分析

Figure 7. Time delay Spearman analysis of rainfall−overall displacement of periodic term

下载: 全尺寸图片幻灯片

图 8 总体周期项位移预测结果

Figure 8. Prediction results of overall displacement of periodic term

下载: 全尺寸图片幻灯片

图 9 不同优化算法下模型的总体周期项位移预测结果

Figure 9. Prediction results of overall displacement of periodic term in optimization models with different algorithms

下载: 全尺寸图片幻灯片

图 10 总体累积位移预测结果

Figure 10. Overall cumulative displacement prediction results

下载: 全尺寸图片幻灯片

图 11 不同优化算法下模型的总体累积位移预测结果

Figure 11. Prediction results of overall cumulative displacement in optimization models with different algorithms

下载: 全尺寸图片幻灯片

表 1 影响因素与总体周期项位移的关联性

Table 1. Correlation between influencing factors and overall displacement of periodic term

影响因子	垂直位移	水平位移	当前月降雨量	前一月累积降雨量	累积降雨量
相关性系数	0.75	0.96	0.82	0.85	0.7

下载: 导出CSV

表 2 不同模型下总体周期项位移的预测精度和预测时间

Table 2. Prediction accuracy and prediction time of overall displacement of periodic term for different models

模型类型	R_RMSE/ mm	M_MAE/ mm	R²	预测时间/s
LSTM	1.55	0.99	0.86	3.20
GRU	1.28	0.95	0.90	4.12
BiLSTM	1.16	0.72	0.91	5.63
CNN-BiLSTM-AM	0.77	0.47	0.96	6.60
SSA-CNN-BiLSTM-AM	0.30	0.58	0.96	6.01
WOA-CNN-BiLSTM-AM	0.37	0.54	0.97	6.23
NGO-CNN-BiLSTM-AM	0.29	0.42	0.97	6.16
NGO-CNN-BiLSTM-AM-未滞后	0.31	0.70	0.95	6.43

下载: 导出CSV

表 3 不同模型下总体累积位移的预测精度

Table 3. Prediction accuracy of overall cumulative displacement under different models

模型类型	R_RMSE/ mm	M_MAE/ mm	R²
LSTM	2.36	1.31	0.90
GRU	1.61	1.09	0.93
BiLSTM	1.18	0.90	0.95
CNN-BiLSTM-AM	0.73	0.60	0.96
SSA-CNN-BiLSTM-AM	0.69	0.56	0.97
WOA-CNN-BiLSTM-AM	0.64	0.52	0.98
NGO-CNN-BiLSTM-AM	0.22	0.37	0.98
NGO-CNN-BiLSTM-AM-未滞后	0.82	0.67	0.97

下载: 导出CSV

参考文献(27)

[1]	邢保印, 张炜怡, 章广成, 等. 基于变形速率分解的阶跃型滑坡预测——以呷爬滑坡为例[J]. 岩石力学与工程学报, 2023, 42(3): 685-697. XING Baoyin, ZHANG Weiyi, ZHANG Guangcheng, et al. Prediction of step-type landslides based on deformation rate decomposition: a case study of Gapa landslide[J]. Chinese Journal of Rock Mechanics and Engineering, 2023, 42(3): 685-697.
[2]	高雅萍, 陈曦, 涂锐. 顾及降雨影响的动态优化时滞时序GM(1, 2)模型在滑坡位移预测中的应用[J]. 测绘学报, 2022, 51(10): 2183-2195. doi: 10.11947/j.AGCS.2022.20220290 GAO Yaping, CHEN Xi, TU Rui. Application of dynamic optimization time-delay GM(1,2) model in landslide displacement prediction considering the influence of rainfall[J]. Acta Geodaetica et Cartographica Sinica, 2022, 51(10): 2183-2195. doi: 10.11947/j.AGCS.2022.20220290
[3]	李骅锦, 许强, 何雨森, 等. “阶跃式” 滑坡位移预测及阈值分析的ARMA-(LASSO-ELM)-Copula模型[J]. 岩石力学与工程学报, 2017, 36(增2): 4075-4084. LI Huajin, XU Qiang, HE Yusen, et al. ARMA-(LASSO-ELM)-Copula model for “step” landslide displacement prediction and threshold analysis[J]. Chinese Journal of Rock Mechanics and Engineering, 2017, 36(S2): 4075-4084.
[4]	DENG L Z, SMITH A, DIXON N, et al. Machine learning prediction of landslide deformation behaviour using acoustic emission and rainfall measurements[J]. Engineering Geology, 2021, 293: 106315 doi: 10.1016/j.enggeo.2021.106315
[5]	NAVA L, CARRARO E, REYES-CARMONA C, et al. Landslide displacement forecasting using deep learning and monitoring data across selected sites[J]. Landslides, 2023, 20(10): 2111-2129. doi: 10.1007/s10346-023-02104-9
[6]	李达, 瞿伟, 张勤, 等. 融合多层感知机和优化支持向量回归的滑坡位移预测模型[J]. 武汉大学学报(信息科学版), 2023, 48(8): 1380-1388. LI Da, QU Wei, ZHANG Qin, et al. Landslide displacement prediction model integrating multi-layer perceptron and optimized support vector regression[J]. Geomatics and Information Science of Wuhan University, 2023, 48(8): 1380-1388.
[7]	杨背背, 殷坤龙, 杜娟. 基于时间序列与长短时记忆网络的滑坡位移动态预测模型[J]. 岩石力学与工程学报, 2018, 37(10): 2334-2343. YANG Beibei, YIN Kunlong, DU Juan. A model for predicting landslide displacement based on time series and long and short term memory neural network[J]. Chinese Journal of Rock Mechanics and Engineering, 2018, 37(10): 2334-2343.
[8]	ZHANG Y G, TANG J, CHENG Y, et al. Prediction of landslide displacement with dynamic features using intelligent approaches[J]. International Journal of Mining Science and Technology, 2022, 32(3): 539-549. doi: 10.1016/j.ijmst.2022.02.004
[9]	SONG K L, YANG H Q, LIANG D, et al. Step-like displacement prediction and failure mechanism analysis of slow-moving reservoir landslide[J]. Journal of Hydrology, 2024, 628: 130588. doi: 10.1016/j.jhydrol.2023.130588
[10]	于群, 霍筱东, 何剑, 等. 基于斯皮尔曼相关系数和系统惯量的中国电网停电事故趋势预测[J]. 中国电机工程学报, 2023, 43(14): 5372-5381. YU Qun, HUO Xiaodong, HE Jian, et al. Trend prediction of power blackout accidents in Chinese power grid based on Spearman’s correlation coefficient and system inertia[J]. Chinese Journal of Electrical Engineering, 2023, 43(14): 5372-5381.
[11]	ZHANG J R, LIN C Y, TANG H M, et al. Input-parameter optimization using a SVR based ensemble model to predict landslide displacements in a reservoir area–A comparative study[J]. Applied Soft Computing, 2024, 150: 111107. doi: 10.1016/j.asoc.2023.111107
[12]	袁维, 孙瑞峰, 钟辉亚, 等. 阶跃型滑坡综合变形预测及监测预警方法研究[J]. 水利学报, 2023, 54(4): 461-473. YUAN Wei, SUN Ruifeng, ZHONG Huiya, et al. Research on comprehensive deformation prediction and monitoring and early warning method for step-like landslide[J]. Journal of Hydraulic Engineering, 2023, 54(4): 461-473.
[13]	王震豪, 聂闻, 许汉华, 等. 基于EEMD-Prophet-LSTM的滑坡位移预测[J]. 中国科学院大学学报, 2023, 40(4): 514-522. WANG Zhenhao, NIE Wen, XU Hanhua, et al. Prediction of landslide displacement based on EEMD-prophet-LSTM[J]. Journal of University of Chinese Academy of Sciences, 2023, 40(4): 514-522.
[14]	XU S L, NIU R Q. Displacement prediction of Baijiabao landslide based on empirical mode decomposition and long short-term memory neural network in Three Gorges area, China[J]. Computers & Geosciences, 2018, 111: 87-96.
[15]	MURAWWAT S, ASIF H M, IJAZ S, et al. Denoising and classification of arrhythmia using MEMD and ANN[J]. Alexandria Engineering Journal, 2022, 61(4): 2807-2823. doi: 10.1016/j.aej.2021.08.014
[16]	MASSAR H, STERGIADIS C, NSIRI B, et al. EMD-BSS: a hybrid methodology combining empirical mode decomposition and blind source separation to eliminate the ocular artifacts from EEG recordings[J]. Biomedical Signal Processing and Control, 2024, 95: 106475. doi: 10.1016/j.bspc.2024.106475
[17]	TARMANINI C, SARMA N, GEZEGIN C, et al. Short term load forecasting based on ARIMA and ANN approaches[J]. Energy Reports, 2023, 9: 550-557.
[18]	LIN H J, ZHANG S R, LI Q H, et al. A new method for heart rate prediction based on LSTM-BiLSTM-Att[J]. Measurement, 2023, 207: 112384. doi: 10.1016/j.measurement.2022.112384
[19]	张振坤, 张冬梅, 李江, 等. 基于多头自注意力机制的LSTM-MH-SA滑坡位移预测模型研究[J]. 岩土力学, 2022, 43(增2): 477-486, 507. ZHANG Zhenkun, ZHANG Dongmei, LI Jiang, et al. Study on LSTM-MH-SA landslide displacement prediction model based on multi-head self-attention mechanism[J]. Rock and Soil Mechanics, 2022, 43(S2): 477-486, 507.
[20]	ZHANG D D, CHEN B A, ZHU H Y, et al. Short-term wind power prediction based on two-layer decomposition and BiTCN-BiLSTM-attention model[J]. Energy, 2023, 285: 128762. doi: 10.1016/j.energy.2023.128762
[21]	NIU D X, YU M, SUN L J, et al. Short-term multi-energy load forecasting for integrated energy systems based on CNN-BiGRU optimized by attention mechanism[J]. Applied Energy, 2022, 313: 118801. doi: 10.1016/j.apenergy.2022.118801
[22]	包艳艳, 杨广泽, 陈伟, 等. 基于SBSS与CNN的750 kV变压器和尖板的放电信号声纹识别[J]. 西南交通大学学报, 2025, 60(3): 781-792. BAO Yanyan, YANG Guangze, CHEN Wei, et al. Voiceprint recognition of discharge aliasing signals from 750 kV transformer and pin-plate based on sparse representation theory and convolutional neural network[J]. Journal of Southwest Jiaotong University, 2025, 60(3): 781-792.
[23]	王瑞, 张璐婷, 逯静. 基于新型相似日选取和VMD-NGO-BiGRU的短期光伏功率预测[J]. 湖南大学学报(自然科学版), 2024, 51(2): 68-80. WANG Rui, ZHANG Luting, LU Jing. Short term photovoltaic power prediction based on new similar day selection and VMD-NGO-BiGRU[J]. Journal of Hunan University (Natural Sciences), 2024, 51(2): 68-80.
[24]	张凯, 张科, 保瑞, 等. 基于优化经验模态分解和聚类分析的滑坡位移智能预测研究[J]. 岩土力学, 2021, 42(1): 211-223. ZHANG Kai, ZHANG Ke, BAO Rui, et al. Intelligent prediction of landslide displacements based on optimized empirical mode decomposition and K-Mean clustering[J]. Rock and Soil Mechanics, 2021, 42(1): 211-223.
[25]	张勤, 白正伟, 黄观文, 等. GNSS滑坡监测预警技术进展[J]. 测绘学报, 2022, 51(10): 1985-2000. doi: 10.11947/j.AGCS.2022.20220299 ZHANG Qin, BAI Zhengwei, HUANG Guanwen, et al. Review of GNSS landslide monitoring and early warning[J]. Acta Geodaetica et Cartographica Sinica, 2022, 51(10): 1985-2000. doi: 10.11947/j.AGCS.2022.20220299
[26]	廖康, 吴益平, 李麟玮, 等. 基于时间序列与GWO-ELM模型的滑坡位移预测[J]. 中南大学学报(自然科学版), 2019, 50(3): 619-626. LIAO Kang, WU Yiping, LI Linwei, et al. Displacement prediction model of landslide based on time series and GWO-ELM[J]. Journal of Central South University (Science and Technology), 2019, 50(3): 619-626.
[27]	张凯翔, 牛瑞卿, 胡友健, 等. 基于小波变换及外因响应的滑坡位移预测[J]. 中国矿业大学学报, 2017, 46(4): 924-931. ZHANG Kaixiang, NIU Ruiqing, HU Youjian, et al. Landslide displacement prediction based on wavelet transform and external cause[J]. Journal of China University of Mining & Technology, 2017, 46(4): 924-931.

施引文献

附加材料(0)

访问统计

点击查看大图

图(11) / 表(3)

计量

文章访问数: 42
HTML全文浏览量: 36
PDF下载量: 7
被引次数: 0

基于NGO优化的CNN-BiLSTM-AM滑坡位移预测模型

doi: 10.3969/j.issn.0258-2724.20240550

作者简介: 王惠琴（1971—），女，教授，博士，研究方向为无线光通信理论与技术，智能信息与多媒体信号处理，E-mail：whq1222@lut.edu.cn

计量

出版历程

NGO-Based CNN-BiLSTM-AM Model for Landslide Displacement Prediction

计量

出版历程

目录

作者简介:
王惠琴（1971—），女，教授，博士，研究方向为无线光通信理论与技术，智能信息与多媒体信号处理，E-mail：whq1222@lut.edu.cn