| CNN-LSTM组合模型在隧道衬砌变形预测中的应用 |
| Application of CNN LSTM Combined Model for Predicting Tunnel Lining Deformation |
| |
| 中文关键词:隧道变形预测 时空特征 CNN-LSTM组合模型 |
| 英文关键词:tunnel deformation prediction spatio temporal characteristics CNN LSTM combined model |
| 基金项目: |
|
| 摘要点击次数: 201 |
| 全文下载次数: 183 |
| 中文摘要: |
| 下穿隧道施工对邻近既有隧道的影响不容忽视，加之传统的预测模型难以捕捉隧道变形数据包含的复杂时空特征，导致拟合效果较差。为解决上述问题，以巫山隧道下穿岳家岭隧道为研究对象，采用CNN (convolutional neural network) LSTM（long short term memory）组合优化模型开展了隧道衬砌变形预测研究，并引入多个统计学指标对模型的预测精度进行了验证。结果表明：CNN LSTM组合模型不仅可以处理包含空间和时间依赖的复杂任务，而且在拟合震荡数据方面具有较强的优势，能较好地捕捉输入数据中的重要局部特征，增强对峰值和峰谷的识别能力。合理的架构设计和优化策略能够充分发挥CNN与LSTM模型的优势，提高预测模型的准确性和鲁棒性。 |
| 英文摘要: |
| The impact of tunnel construction on adjacent existing tunnels is significant, and traditional prediction models often struggle to capture the complex spatio temporal characteristics of tunnel deformation data, resulting in suboptimal fitting performance. To address these issues, this study investigates the Wushan Tunnel undercrossing the Yuejialing Tunnel, utilizing a combined CNN (Convolutional Neural Network) LSTM (Long Short Term Memory) model to predict tunnel lining deformation. Several statistical indicators are introduced to validate the prediction accuracy of the model. The results indicate that the CNN LSTM combined model can effectively work with datas involving both spatial and temporal dependencies, demonstrating a strong capability to fit oscillatory data and capture critical local features. A well designed and optimized architecture leverages the strengths of both CNN and LSTM models, significantly enhancing the accuracy and robustness of the prediction model. |
| 李成辉.CNN-LSTM组合模型在隧道衬砌变形预测中的应用[J].国防交通工程与技术,2024,22(6):35~40 |
| 查看全文 查看/发表评论 下载PDF阅读器 |
|
| 关闭 |
