We are thrilled to congratulate Professor Shoukai Chen and his research team from the School of Water Conservancy at North China University of Water Resources and Electric Power (NCWU) on theirinnovative publication in the international journal Evidence in Earth Science (EIES). Their article, titled “Prediction of Dike Seepage Pressure Based on ISSA-BiLSTM”, introduces a novel model that significantly advances dam safety monitoring and seepage prediction.
Key Research Highlights
The study addresses critical challenges in traditional dam seepage pressure prediction models, such as susceptibility to local optima and low efficiency. The team proposed an Improved Sparrow Search Algorithm (ISSA) enhanced by nonlinear Sine Cosine optimization and adaptive producer-scrounger ratio adjustment. This algorithm was synergistically integrated with a Bidirectional Long Short-Term Memory (BiLSTM) neural network to develop the ISSA-BiLSTM prediction model.
By leveraging LightGBM technology for feature dimensionality reduction, the team identified seven key influencing factors—including upstream water levels, rainfall, and temperature components—as optimal inputs for the model. The ISSA-BiLSTM demonstrated remarkable performance, achieving an R² of 0.987 and reducing prediction errors by 30% and 20% compared to standalone BiLSTM and SSA-BiLSTM models, respectively.
Practical Significance
The research holds profound implications for dam safety management, particularly as aging infrastructure and extreme weather events escalate risks globally. The ISSA-BiLSTM model provides a robust tool for real-time seepage monitoring, enabling early hazard detection and informed decision-making in water resource projects.
Acknowledgments and Funding
This work was supported by the National Natural Science Foundation of China (Grant No. 51979169) and the Henan Province Innovation Talent Support Plan (24HASTIT017). The authors also acknowledge NCWU for its institutional support.
Professor Chen remarked, “Our model bridges the gap between theoretical optimization and practical application in dam safety. We aim to further refine its accuracy and adaptability for global infrastructure challenges.”
The full article is available via DOI: 10.63221/eies.v1i01.1-16.