Intelligent Structural Health Monitoring of Hydraulic Structures Using Hybrid Deep Learning Models

Hydraulic structures such as dams, spillways, and energy dissipation systems are critical components of water resource infrastructure and are continuously exposed to complex hydraulic, mechanical, and environmental loads. These conditions often lead to progressive deterioration that is difficult to identify using conventional Structural Health Monitoring (SHM) techniques, particularly when large-scale, heterogeneous, and time-dependent sensor data are involved. This study presents a hybrid physics-informed deep learning framework for advanced health monitoring and damage assessment of hydraulic structures. The proposed approach integrates Convolutional Neural Networks (CNNs) to extract spatial features from vibration, strain, and pressure measurements with Long Short-Term Memory (LSTM) networks to capture the temporal evolution of structural responses under varying operational conditions. To ensure physically consistent predictions, physics-based constraints derived from hydraulic and structural mechanics are incorporated into the learning process. The model is evaluated using a combination of high-resolution numerical simulations and long-term monitoring data, representing both normal and degraded structural states. The results demonstrate that the proposed CNN-LSTM framework significantly outperforms conventional machine learning models in terms of anomaly detection accuracy, damage characterization, and remaining useful life estimation. The findings highlight the strong potential of the proposed methodology for real-time structural assessment and predictive maintenance, offering a reliable decision-support tool for improving the safety, resilience, and operational efficiency of critical hydraulic infrastructure.

کلیدواژه ها:

Hydraulic Structures ، Structural Health Monitoring (SHM) ، Hybrid Deep Learning ، CNN-LSTM Architecture ، Predictive Maintenance