Physics-Informed Neural Network Modeling of Biodegradable Contaminant Transport through GCL/SL Composite Liners

2026-06-03 · Source: Takara TLDR - Daily AI Papers · Field: Science & Research — Artificial Intelligence & Machine Learning, Engineering & Applied Sciences, Environmental Science & Earth Systems · Depth: Expert, quick

Summary

A two-domain physics-informed neural network (PINN) framework has been developed to model biodegradable contaminant transport through GCL/SL composite liner systems. This framework treats the thin GCL layer using a steady-state advection-dispersion-biodegradation formulation, while the underlying soil liner is modeled as a transient transport domain. Researchers evaluated two PINN formulations: a standard PINN (Std-PINN) with soft constraint enforcement and a hard-constrained PINN (H-PINN) where boundary and initial conditions are directly embedded. The H-PINN demonstrated superior accuracy and stability, reducing the Mean Absolute Error (MAE) from approximately 0.058-0.067 for Std-PINN to 0.011-0.023, and the Mean Relative Error (MRE) from 9.10%-19.16% to 2.08%-3.14%. Parametric analyses confirmed the H-PINN with a "tanh" activation function and an optimized network structure provided the best predictive accuracy. The H-PINN was also extended for inverse modeling, reliably identifying SL degradation half-life from limited observations.

Key takeaway

For environmental engineers or research scientists modeling contaminant transport through composite liners, you should prioritize hard-constrained Physics-Informed Neural Networks (H-PINNs). This approach significantly improves prediction accuracy and stability, especially during early transport stages and under high leachate heads. Implementing H-PINNs with a "tanh" activation function can reduce Mean Absolute Error to 0.011-0.023. This offers more reliable simulations for design and risk assessment. Consider extending H-PINNs for inverse modeling to identify critical material properties like degradation half-life.

Key insights

Hard-constrained PINNs significantly improve accuracy and stability for modeling contaminant transport in composite liners.

Principles

• Embedding constraints directly enhances PINN performance.
• Advective transport increases early-stage modeling errors.

Method

A two-domain PINN framework models GCL as steady-state and soil liner as transient, evaluating standard vs. hard-constrained approaches for contaminant transport.

In practice

• Use H-PINN for higher accuracy in transport modeling.
• Employ "tanh" activation for optimal PINN prediction.
• Apply H-PINN for inverse parameter identification.

Topics

• Physics-Informed Neural Networks
• Contaminant Transport Modeling
• GCL/SL Composite Liners
• Inverse Modeling
• Biodegradation
• Geotechnical Engineering

Best for: AI Scientist, Research Scientist

Related on AIssential

• Adaptive Domain Decomposition Physics-Informed Neural Networks for Traffic State Estimation with Sparse Sensor Data — ADD-PINN improves traffic state estimation from sparse sensor data by adaptively decomposing domains based on residual errors.
• CO$_2$ sequestration hybrid solver using isogeometric alternating-directions and collocation-based robust variational physics informed neural networks (IGA-ADS-CRVPINN) — Hybridizing IGA-ADS with CRVPINN significantly accelerates $CO_{2}$ sequestration simulations while maintaining accuracy.
• PINN loss functions: why physics-informed networks often fail to train — PINNs struggle to train due to severe gradient imbalance and scale differences among their multi-term loss functions.
• Material-Agnostic Zero-Shot Thermal Inference for Metal Additive Manufacturing via a Parametric PINN Framework — A parametric PINN framework enables zero-shot thermal modeling across diverse materials in metal AM with high efficiency.
• Using AI, Mathematicians Find Hidden Glitches in Fluid Equations — Bespoke PINNs can precisely identify previously unobserved unstable fluid singularity candidates across diverse fluid models.
• Physics-informed convolutional neural networks for fluid flow through porous media — Physics-informed CNNs accurately predict fluid flow in porous media, significantly accelerating traditional LBM simulations.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Takara TLDR - Daily AI Papers.