A general progressive decomposition long-term prediction network model for high-speed railway bridge pier settlement

doi:10.11947/j.AGCS.2024.20230387

Abstract

Abstract:

Uneven settlement of high-speed railway bridge pier is one of the potential causes leading to track irregularities. Accurately predicting settlement of bridge pier is of significant importance for ensuring the reliability and safety of railway construction and operation. Most conventional time series prediction models are tested only on well-preprocessed datasets without missing values. However, in real-world scenarios of high-speed railway bridge pier settlement, the data are characterized by infrequent and irregular observation intervals and complex, variable settlement patterns, posing challenges for long-term prediction. To address this, we introduce the general progressive decomposition long-term prediction network (GPDLPnet), which abandons traditional preprocessing concepts and embeds the preprocessing phase within the network structure, achieving progressive preprocessing during training. In each iteration, GPDLPnet uses an improved diagonally-masked self-attention (IDMSA) module to analyze missing patterns in the settlement data, then decomposes and reconstructs the data into high-frequency, low-frequency, and trend sub-components through an improved complete ensemble empirical mode decomposition with adaptive noise (ICEEMDAN) module. These sub-components serve as feature inputs for the BiLSTM-RSA-Resnet prediction module, which outputs recursive predictions, thus enabling long-term prediction of high-speed railway bridge pier settlement. Utilizing real-world engineering data, experiments under two typical observation modes, high-frequency and low-frequency, are conducted. GPDLPnet demonstrates excellent predictive performance over a 3-4 month, surpassing seven other models in accuracy indexes.

Key words: deep learning, high-speed railway bridge pier, settlement prediction, residual network, convolutional neural network

CLC Number:

P258

Xunqiang GONG, Hongyu WANG, Tieding LU, Wei YOU. A general progressive decomposition long-term prediction network model for high-speed railway bridge pier settlement[J]. Acta Geodaetica et Cartographica Sinica, 2024, 53(6): 1113-1127.

Figures/Tables 15

Fig.1

Fig.2

Fig.3

Fig.4

Fig.5

Fig.6

Fig.7

Fig.8

Fig.9

Tab.1

Fig.10

Fig.11

Fig.12

Fig.13

Tab.2

References 35

[1]	黎奇, 白征东, 陈波波, 等. GNSS/INS多传感器组合高速铁路轨道测量系统[J]. 测绘学报, 2020, 49(5):569-579. DOI:10.11947/j.AGCS.2020.20190344.
	LI Qi, BAI Zhengdong, CHEN Bobo, et al. A novel track measurement system based on GNSS/INS and multisensor for high-speed railway[J]. Acta Geodaetica et Cartographica Sinica, 2020, 49(5):569-579. DOI:10.11947/j.AGCS.2020.20190344.
[2]	中国铁路总公司. 铁路工程沉降变形观测与评估技术规程Q/CR 9230—2016[S]. 北京: 中国铁道出版社, 2016.
	China Railway Corporation. Observation and evaluation specification for settlement deformation of railway engineering Q/CR 9230—2016[S]. Beijing: China Railway Publishing House, 2016.
[3]	张献州, 夏晨翕, 陈霄, 等. IGM-FM串联模型在高铁路基沉降预测中的应用[J]. 重庆交通大学学报(自然科学版), 2020, 39(11):99-108.
	ZHANG Xianzhou, XIA Chenxi, CHEN Xiao, et al. Application of series IGM-FM combination model in prediction of subgrade settlement of high-speed railway[J]. Journal of Chongqing Jiaotong University (Natural Science), 2020, 39(11):99-108.
[4]	GONG X, LI Z. Bridge pier settlement prediction in high-speed railway via autoregressive model based on robust weighted total least-squares[J]. Survey Review, 2018, 50(359):147-154.
[5]	GONG Xunqiang, WANG Hongyu, LU Tieding, et al. Combined prediction model for high-speed railway bridge pier settlement based on robust weighted total least-squares autoregression and adaptive dynamic cubic exponential smoothing[J]. Journal of Surveying Engineering, 2023, 149(2):04023001.
[6]	王秀英, 雷可, 王滕, 等. 洞桩法车站下穿既有管线工程的沉降预测[J]. 土木工程学报, 2021, 54(S1):65-75.
	WANG Xiuying, LEI Ke, WANG Teng, et al. Settlement prediction of tunnel pile method station under existing pipeline project[J]. China Civil Engineering Journal, 2021, 54(S1):65-75.
[7]	陈军, 刘万增, 武昊, 等. 智能化测绘的基本问题与发展方向[J]. 测绘学报, 2021, 50(8):995-1005. DOI:10.11947/j.AGCS.2021.20210235.
	CHEN Jun, LIU Wanzeng, WU Hao, et al. Smart surveying and mapping: fundamental issues and research agenda[J]. Acta Geodaetica et Cartographica Sinica, 2021, 50(8):995-1005. DOI:10.11947/j.AGCS.2021.20210235.
[8]	陈军, 王艳慧, 武昊, 等. 时空信息赋能高质量发展的基本问题与发展方向[J]. 时空信息学报, 2023, 30(1):1-11.
	CHEN Jun, WANG Yanhui, WU Hao, et al. Basic issues and development directions of high-quality development empowered by spatio-temporal information[J]. Journal of Spatio-Temporal Information, 2023, 30(1):1-11.
[9]	郭澳庆, 胡俊, 郑万基, 等. 时序InSAR滑坡形变监测与预测的N-BEATS深度学习法：以新铺滑坡为例[J]. 测绘学报, 2022, 51(10):2171-2182. DOI:10.11947/j.AGCS.2022.20220298.
	GUO Aoqing, HU Jun, ZHENG Wanji, et al. N-BEATS deep learning method for landslide deformation monitoring and prediction based on InSAR: a case study of Xinpu landslide[J]. Acta Geodaetica et Cartographica Sinica, 2022, 51(10):2171-2182. DOI:10.11947/j.AGCS.2022.20220298.
[10]	禹文豪, 魏铖, 陈佳鑫. 面向不平衡POI类别分布的电子地图多层次标签预测模型[J]. 测绘学报, 2023, 52(2):318-328. DOI:10.11947/j.AGCS.2023.20210451.
	YU Wenhao, WEI Cheng, CHEN Jiaxin. Predicting the unbalanced labels of POIs on digital maps using hierarchical model[J]. Acta Geodaetica et Cartographica Sinica, 2023, 52(2):318-328. DOI:10.11947/j.AGCS.2023.20210451.
[11]	MACKENZIE J, RODDICK J F, ZITO R. An evaluation of HTM and LSTM for short-term arterial traffic flow prediction[J]. IEEE Transactions on Intelligent Transportation Systems, 2019, 20(5):1847-1857.
[12]	冯非凡, 武雪玲, 牛瑞卿, 等. 一种V/S和LSTM结合的滑坡变形分析方法[J]. 武汉大学学报(信息科学版), 2019, 44(5):784-790.
	FENG Feifan, WU Xueling, NIU Ruiqing, et al. A landslide deformation analysis method using V/S and LSTM[J]. Geomatics and Information Science of Wuhan University, 2019, 44(5):784-790.
[13]	ZHENG Haifeng, LIN Feng, FENG Xinxin, et al. A hybrid deep learning model with attention-based conv-LSTM networks for short-term traffic flow prediction[J]. IEEE Transactions on Intelligent Transportation Systems, 2021, 22(11):6910-6920.
[14]	周于涛, 吴华意, 成洪权, 等. 结合自注意力机制和结伴行为特征的行人轨迹预测模型[J]. 武汉大学学报(信息科学版), 2020, 45(12):1989-1996.
	ZHOU Yutao, WU Huayi, CHENG Hongquan, et al. Pedestrian trajectory prediction model based on self-attention mechanism and group behavior characteristics[J]. Geomatics and Information Science of Wuhan University, 2020, 45(12):1989-1996.
[15]	王利, 许豪, 舒宝, 等. 利用互信息和IPSO-LSTM进行滑坡监测多源数据融合[J]. 武汉大学学报(信息科学版), 2021, 46(10):1478-1488.
	WANG Li, XU Hao, SHU Bao, et al. A multi-source heterogeneous data fusion method for landslide monitoring with mutual information and IPSO-LSTM neural network[J]. Geomatics and Information Science of Wuhan University, 2021, 46(10):1478-1488.
[16]	LI Guannan, ZHAO Xiaowei, FAN Cheng, et al. Assessment of long short-term memory and its modifications for enhanced short-term building energy predictions[J]. Journal of Building Engineering, 2021, 43:103182.
[17]	陈军, 艾廷华, 闫利, 等. 智能化测绘的混合计算范式与方法研究[J/OL]. 测绘学报: 1-19 [2024-05-20].http://kns.cnki.net/kcms/detail/11.2089.P.20240415.1049.002.html.
	CHEN Jun, AI Tinhua, YAN Li, et al. Hybrid computational paradigm and methods for intelligentized surveying and mapping[J/OL]. Acta Geodaetica et Cartographica Sinica: 1-19 [2024-05-20].http://kns.cnki.net/kcms/detail/11.2089.P.20240415.1049.002.html.
[18]	CHEN X, SUN L. Bayesian temporal factorization for multidimensional time series prediction[J]. IEEE Trans Pattern Anal Mach Intell, 2022, 44(9):4659-4673.
[19]	PURWINS H, LI Bo, VIRTANEN T, et al. Deep learning for audio signal processing[J]. IEEE Journal of Selected Topics in Signal Processing, 2019, 13(2):206-219.
[20]	李新琴, 史天运, 李平, 等. 基于文本的高速铁路信号设备故障知识抽取方法研究[J]. 铁道学报, 2021, 43(3):92-100.
	LI Xinqin, SHI Tianyun, LI Ping, et al. Research on knowledge extraction method for high-speed railway signal equipment fault based on text[J]. Journal of the China Railway Society, 2021, 43(3):92-100.
[21]	DU Wenjie, CÔTÉ D, LIU Yan. SAITS: self-attention-based imputation for time series[J]. Expert Systems with Applications, 2023, 219:119619.
[22]	HUANG Jin, ZHANG Tinghua, ZHU Jia, et al. A deep embedding model for knowledge graph completion based on attention mechanism[J]. Neural Computing and Applications, 2021, 33(15):9751-9760.
[23]	罗亦泳, 黄城, 张静影. 基于变分模态分解的变形监测数据去噪方法[J]. 武汉大学学报(信息科学版), 2020, 45(5):784-790.
	LUO Yiyong, HUANG Cheng, ZHANG Jingying. Denoising method of deformation monitoring data based on variational mode decomposition[J]. Geomatics and Information Science of Wuhan University, 2020, 45(5):784-790.
[24]	邢世玲, 吕双双, 朱利明, 等. 基于ICEEMDAN的连续梁桥车致振动信号的HHT分析[J]. 西南交通大学学报, 2021, 56(3):477-484, 492.
	XING Shiling, LÜ Shuangshuang, ZHU Liming, et al. Hilbert-Huang transfer analysis on vehicle-induced vibration signal of continuous bridges based on ICEEMDAN[J]. Journal of Southwest Jiaotong University, 2021, 56(3):477-484, 492.
[25]	程肇兰, 张小强, 梁越. 基于LSTM网络的铁路货运量预测[J]. 铁道学报, 2020, 42(11):15-21.
	CHENG Zhaolan, ZHANG Xiaoqiang, LIANG Yue. Railway freight volume prediction based on LSTM network[J]. Journal of the China Railway Society, 2020, 42(11):15-21.
[26]	YANG C, KIM T, WANG R, et al. Show, attend, and translate: unsupervised image translation with self-regularization and attention[J]. IEEE Trans Image Process, 2019, 28(10):4845-4856.
[27]	张淑清, 杜灵韵, 王册浩, 等. 基于格拉姆角场与改进CNN-ResNet的风电功率预测方法[J]. 电网技术, 2023, 47(4):1540-1547.
	ZHANG Shuqing, DU Lingyun, WANG Cehao, et al. Wind power forecasting method based on GAF and improved CNN-ResNet[J]. Power System Technology, 2023, 47(4):1540-1547.
[28]	CAO Pei, ZHANG Shengli, TANG Jiong. Preprocessing-free gear fault diagnosis using small datasets with deep convolutional neural network-based transfer learning[J]. IEEE Access, 2018, 6:26241-26253.
[29]	ZHANG X, LIN T, XU J, et al. DeepSpectra: an end-to-end deep learning approach for quantitative spectral analysis[J]. Anal Chim Acta, 2019, 1058:48-57.
[30]	LIU Hui, MI Xiwei, LI Yanfei, et al. Smart wind speed deep learning based multi-step forecasting model using singular spectrum analysis, convolutional gated recurrent unit network and support vector regression[J]. Renewable Energy, 2019, 143:842-854.
[31]	LIVIERIS I E, PINTELAS E, PINTELAS P. A CNN-LSTM model for gold price time-series forecasting[J]. Neural Computing and Applications, 2020, 32(23):17351-17360.
[32]	BARRA S, CARTA S, CORRIGA A, et al. Deep learning and time series-to-image encoding for financial forecasting[J]. IEEE/CAA Journal of Automatics Sinica, 2020, 7(3):683-692.
[33]	PAN Guangliang, WU Qihui, DING Guoru, et al. An autoformer-CSA approach for long-term spectrum prediction[J]. IEEE Wireless Communications Letters, 2023, 12(10):1647-1651.
[34]	龚循强, 汪宏宇, 鲁铁定, 等. 基于RLM-MC的高速铁路桥墩沉降曲线预测方法[J]. 铁道学报, 2022, 44(10):142-149.
	GONG Xunqiang, WANG Hongyu, LU Tieding, et al. A settlement curve prediction method of high-speed railway bridge pier based on RLM-MC[J]. Journal of the China Railway Society, 2022, 44(10):142-149.
[35]	汪宏宇, 龚循强, 鲁铁定, 等. 基于马尔科夫链的曲线拟合法在尾矿坝沉降预测中的应用[J]. 东华理工大学学报(自然科学版), 2021, 44(4):364-369.
	WANG Hongyu, GONG Xunqiang, LU Tieding, et al. Application of curve fitting method based on Markov chain in tailings dam settlement prediction[J]. Journal of East China University of Technology (Natural Science), 2021, 44(4):364-369.

模型	A1		A2		B1		B2
模型	MAE	RMSE	MAE	RMSE	MAE	RMSE	MAE	RMSE
SVR	1.267	1.302	1.365	1.414	1.126	1.196	1.327	1.353
LSTM	1.185	1.212	1.240	1.550	1.191	1.229	1.361	1.382
CNN-LSTM	1.213	1.247	1.155	1.188	1.203	1.248	1.148	1.171
CNN-Attention-LSTM	1.061	1.095	0.813	0.905	1.145	1.192	1.083	1.107
BiLSTM	1.002	1.037	1.092	1.119	1.552	1.576	1.422	1.439
CNN-BiLSTM	1.481	1.513	1.021	1.083	1.045	1.091	1.262	1.282
Resnet	0.597	0.696	1.245	1.307	1.031	1.096	0.555	0.678
GPDLPnet	0.229	0.267	0.355	0.432	0.377	0.467	0.391	0.464

测点	模型	SVR	LSTM	CNN-LSTM	CNN-Attention-LSTM	BiLSTM	CNN-BiLSTM	Resnet	GPDLPnet
E1	MAE	1.144	0.936	0.892	0.772	0.632	0.806	0.836	0.273
E1	RMSE	1.244	0.955	0.933	0.830	0.690	0.849	1.037	0.309
E2	MAE	1.412	1.032	0.772	0.780	0.349	0.748	0.934	0.247
E2	RMSE	1.613	1.050	0.811	0.818	0.453	0.767	1.092	0.355
F1	MAE	0.989	0.603	0.671	0.683	0.504	0.574	0.589	0.278
F1	RMSE	1.288	0.687	0.816	0.809	0.571	0.678	0.774	0.342
F2	MAE	0.996	0.987	0.932	0.948	0.939	0.648	0.641	0.259
F2	RMSE	1.163	1.154	1.061	1.106	1.099	0.768	0.764	0.335
G1	MAE	0.459	0.675	0.408	0.337	0.381	0.350	0.437	0.157
G1	RMSE	0.571	0.746	0.471	0.433	0.457	0.439	0.506	0.242
G2	MAE	0.615	0.482	0.353	0.271	0.774	0.345	0.673	0.233
G2	RMSE	0.692	0.538	0.395	0.304	0.873	0.372	0.821	0.263
H1	MAE	0.785	1.117	1.175	0.938	1.163	1.023	0.639	0.423
H1	RMSE	0.953	1.205	1.278	1.008	1.243	1.106	0.823	0.533
H2	MAE	0.663	0.803	0.855	0.902	0.907	0.875	0.315	0.204
H2	RMSE	0.736	0.918	1.001	1.049	1.025	1.009	0.401	0.243