Physics Informed Neural Network (PINN) enabled Predictive Resilience Framework for Maritime and Multimodal Levee Infrastructure

The performance and resilience of levee systems are governed by complex hydro-mechanical interactions influenced by transient seepage, soil stratification, and environmental loading. Conventional monitoring approaches, while effective in capturing field conditions, lack predictive capability and often fail to integrate subsurface characterization with real-time system response. This study proposes a Physics-Informed Neural Network (PINN) enabled predictive resilience framework for maritime and multimodal levee infrastructure, integrating multi-source sensing, geophysical imaging, and physics-based modeling. The framework leverages Internet of Things (IoT) based sensor networks, including IMU derived tilt and displacement measurements, and environmental variables such as rainfall, temperature, and soil moisture. To enhance subsurface characterization, Electrical Resistivity Imaging (ERI) and Multichannel Analysis of Surface Waves (MASW) are incorporated to capture spatial variability in moisture distribution, stiffness profiles, and potential seepage zones. These datasets are fused with UAV based LiDAR point cloud models to develop high-resolution, temporal geospatial conditional representations of levee geometry and deformation. The integrated dataset is utilized to calibrate finite element method (FEM) based seepage and stability models, enabling accurate representation of coupled hydromechanical behavior. The PINN architecture embeds governing equations of transient flow and unsaturated soil mechanics into the learning process, allowing physically consistent prediction of pore pressure, volumetric moisture content, and deformation fields. A hybrid physics-guided, data-driven digital twin will be developed to continuously assimilate field and geophysical data, providing real-time predictions and identifying anomaly thresholds indicative of instability. The proposed framework advances geotechnical asset management by enabling predictive failure assessment, risk-informed decision-making, and proactive maintenance strategies, thereby enhancing the resilience of critical maritime and multimodal infrastructure systems under extreme environmental conditions.

Language

• English

Project

• Status: Active
• Funding: $82,500.00
• Contract Numbers:

  69A3552348331

• Sponsor Organizations:

  Office of the Assistant Secretary for Research and Technology

  University Transportation Centers Program
  Department of Transportation
  Washington, DC  United States  20590
• Managing Organizations:

  Maritime Transportation Research and Education Center (MarTREC)

  University of Arkansas
  4190 Bell Engineering Center
  Fayetteville, AR  United States  72701
• Performing Organizations:

  Jackson State University, Jackson

  Department of Civil and Environmental Engineering
  Jackson, MS  United States  39217-0168
• Principal Investigators:

  Khan, Sadik

• Start Date: 20260701
• Expected Completion Date: 20270630
• Actual Completion Date: 0
• USDOT Program: University Transportation Centers

Subject/Index Terms

• TRT Terms: Data fusion; Deformation; Disaster resilience; Failure analysis; Geospatial data; Hydromechanics; Levees; Neural networks; Predictive models; Seepage
• Geographic Terms: Jackson (Mississippi)
• Subject Areas: Bridges and other structures; Data and Information Technology; Geotechnology; Highways; Hydraulics and Hydrology; Maintenance and Preservation; Planning and Forecasting;

Filing Info

• Accession Number: 01996311
• Record Type: Research project
• Source Agency: Maritime Transportation Research and Education Center (MarTREC)
• Contract Numbers: 69A3552348331
• Files: UTC, RIP
• Created Date: Jul 21 2026 4:41PM