围海造陆区地面沉降与水文地质变化的耦合预测模型研究

打开文本图片集

DOI：10.19981/j.CN23-1581/G3.2026.12.009

中图分类号：P756.8

文献标志码： A

文章编号：2095-2945（2026）12-0033-04

Abstract： Land subsidence in reclaimed land areas is closely related to hydrogeological changes， affecting environmental security and engineering construction. Traditional prediction methods rely on simple physical models or statistical analysis， making it difficult to capture the complex interaction between settlement and hydrogeological changes. This paper proposes a prediction model based on the coupling of deep neural networks and multiple physical fields. It combines remote sensing data， surface monitoring and hydrogeological information， and improves prediction precision by integrating spatio-temporal characteristics and nonlinear relationships. The model combines Convolutional Neural Networks （CNN） and Long Short-Term Memory networks （LSTM） to accurately simulate the feedback effects between ground subsidence and groundwater flow， demonstrating good adaptability， especially under complex geological conditions. Verification results of the model in multiple experiments show that it can efficiently and accurately predict settlement changes and provide strong support for engineering decisions.

Keywords： reclaimed land area; land subsidence; hydrogeology; deep learning; multi-physical field coupling

地面沉降是全球多个地区面临的重要环境与工程问题[1]，特别是在围海造陆、地下水开采及城市化快速发展的区域[2]。（剩余5623字）