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

Researchers have developed two Physics-Informed Neural Network (PINN) frameworks, one for contaminant transport through composite liner systems and another for microbial interaction modelling, incorporating auxiliary knowledge sources to improve parameter discovery and ecological insights.

A two-domain PINN framework has been developed for contaminant transport through a Geosynthetic Clay Liner (GCL) and Soil Liner (SL) composite liner system. The 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^[1]. Two formulations were evaluated: a standard PINN with soft constraint enforcement (Std-PINN) and a hard-constrained PINN (H-PINN). The H-PINN reduces the optimization burden and provides more accurate and stable concentration predictions. It was further extended to inverse modeling for identifying the SL degradation half-life from limited concentration observations. Separately, a new framework for microbial interaction modelling incorporates auxiliary knowledge sources, such as text and network structure, to improve parameter discovery and ecological insights. This framework is based on a PINN approach and integrates peer-reviewed metagenomics literature to enrich the generalized Lotka-Volterra (gLV) parameters. It also incorporates network-based structural knowledge by explicitly modelling microbial interactions, improving accuracy by up to 53% without knowledge addition^[2].