Multi-Layer echo kernel state network based inland waterway water levels forecasting

doi:10.16411/j.cnki.issn1006-7736.2025.03.007

PDF(7310 KB)

Journal of Dalian Maritime University ›› 2025, Vol. 51 ›› Issue (3) : 64-73. DOI: 10.16411/j.cnki.issn1006-7736.2025.03.007

Multi-Layer echo kernel state network based inland waterway water levels forecasting

Author information +

History +

Abstract

Aiming at the problem that traditional water level forecasting models are difficult to effectively capture complex nonlinear dynamics features and affect prediction accuracy,a water level forecasting model based on multi-layer echo kernel state network (ML-EKSN) was proposed. Gaussian kernel functions was used to achieve high-dimensional nonlinear mapping, and hierarchical feature extraction was carried out through multi-layer structure, significantly improving the dynamic modeling ability of the model. Based on the measured data from 7 stations in the Songhua River Basin, ML-EKSN was used to predict the future 7 step water level characteristics based on 30 step historical data.The r esults show that the performance of ML-EKSN is superior to mainstream models such as LSTM, GRU, RVFL, ESN, EKSN,and Transformer, and can provide an efficient and stable solution for predicting water levels in inland waterways and has good practical application value.

Key words

inland waterways / water level prediction / echo state network (ESN) / Gaussian kernel function / multi-layer neural network

Cite this article

EndNote

Ris (Procite)

Bibtex

Download Citations

Multi-Layer echo kernel state network based inland waterway water levels forecasting[J]. Journal of Dalian Maritime University. 2025, 51(3): 64-73 https://doi.org/10.16411/j.cnki.issn1006-7736.2025.03.007

References

[1] KOVVURI A R, UPPALAPATI P J, BONTHU S, et al. Water level forecasting in reservoirs using time series analysis–auto ARIMA model[C]//Proceedings of the International Conference on Cognitive Computing and Cyber Physical Systems.[S.l.:s.n.],2022.

[2] 王岩, 胡雪峰, 贺海飞. 基于FEFLOW和ARIMA模型的中关铁矿地下水动态模拟预测 [J]. 水资源开发与管理, 2024, 10(9): 60-68.

WANG Y, HU X F,HE H F. Dynamic simulation and prediction of groundwater at Zhongguan iron mine based on FEFLOW and ARIMA models [J]. Water Resources Development and Management, 10(09), 60-68.（in Chinese）

[3] VICCIONE G, GUARNACCIA C, MANCINI S, et al. On the use of ARIMA models for short-term water tank levels forecasting[J]. Water Supply, 2020, 20(3): 787-799.

[4] 王迎飞, 黄应平, 肖敏, 等. 基于注意力机制的LSTM长江汛期水位预测方法研究 [J]. 三峡大学学报(自然科学版), 2022, 44(3): 13-19.

WANG Y F, HUANG Y P, XIAO M, et al. Research on an LSTM-based water level prediction method for the Yangtze River flood season using an attention mechanism [J]. Journal of Three Gorges University (Natural Science Edition), 44(3), 13-19. （in Chinese）

[5] 潘明阳, 周海南, 李增辉, 等. 智能水位预测服务系统研究 [J]. 大连海事大学学报, 2020, 46(3): 31-37.

PAN M Y, ZHOU H N, LI Z H, et al.Research on an intelligent water level forecasting service system[J]. Journal of Dalian Maritime University, 46(3):31-37. （in Chinese）

[6] ONSAY E A, BULAO R J G, RABAJANTE J F. Bagyong Kristine (TS Trami) in bicol, Philippines: flood risk forecasting, disaster risk preparedness predictions and lived experiences through machine learning (ML), econometrics, and hermeneutic analysis [J]. Natural Hazards Research, 2025.https://sciencedirect.com/science/article/pii/S2666592125000174.

[7] LIU D, JIANG W, MU L, et al. Streamflow prediction using deep learning neural network: case study of Yangtze River [J]. IEEE access, 2020, 8: 90069-90086.

[8] GHORBANI M A, DEO R C, KARIMI V, et al. Implementation of a hybrid MLP-FFA model for water level prediction of Lake Egirdir, Turkey [J]. Stochastic Environmental Research and Risk Assessment, 2018, 32: 1683-1697.

[9] CHO M, KIM C, JUNG K, et al. Water level prediction model applying a long short-term memory (LSTM)–gated recurrent unit (GRU) method for flood prediction [J]. Water, 2022, 14(14): 2221.

[10] DOSOVITSKIY A, BEYER L, KOLESNIKOV A, et al. An image is worth 16x16 words: transformers for image recognition at scale [J]. arXiv preprint,2020: arXiv:201011929.

[11] YANG C, WANG D, TANG J, et al. Multi-reservoir ESN-based prediction strategy for dynamic multi-objective optimization [J]. Information Sciences, 2024, 652: 119495.

[12] KIM T, KING B R. Time series prediction using deep echo state networks [J]. Neural Computing and Applications, 2020, 32(23): 17769-17787.

[13] SHI L, LU R, LIU Z, et al. An improved robust kernel adaptive filtering method for time-series prediction [J]. IEEE Sensors Journal, 2023, 23(18): 21463-21473.

[14] LI X, BI F, ZHANG L, et al. Rotating machinery faults detection method based on deep echo state network [J]. Applied Soft Computing, 2022, 127: 109335.

[15] TANG G X, YU L,LV W Y, et al. Dual-kernel echo state network for nonlinear time series prediction[J]. Proceeding of the Romanian Academy, Series A:Mathematics Physics Technical Sciences Information Science, 2023, 24(2): 179-190.

[16] SHI L, LU R, LIU Z, et al. An improved robust kernel adaptive filtering method for time-series prediction[J]. IEEE Sensors Journal, 2023, 23(18): 21463-21473.

[17] LIU Z, XU X H, PAN M, et al. Weighted error-output recurrent echo kernel state network for multi-step water level prediction[J]. Applied Soft Computing, 2023, 137: 110131.

[18] JAEGER H. The “echo state” approach to analysing and training recurrent neural networks-with an erratum note [J]. Bonn, Germany: German National Research Center for Information Technology GMD Technical Report, 2001, 148(34): 13.

[19] ADNAN R, MOSTAFA R, ISLAM A, et al. Improving drought modeling using hybrid random vector functional link methods. Water,2021, 13(23):3379.

[20] LI Y S, LIU S, ZHANG Y, et al. A ladder water level prediction model for the Yangtze River based on transfer learning and transformer [J]. IEEE Transactions on Geoscience and Remote Sensing, 2024,62:4211511.