Physics-Informed Neural Network Modeling of Biodegradable Contaminant Transport through GCL/SL Composite Liners
Researchers have developed a novel two-domain physics-informed neural network (PINN) framework to model contaminant transport through composite liner systems. This framework utilizes a hard-constrained PINN (H-PINN) approach, which embeds boundary and initial conditions directly into the network, leading to significantly more accurate and stable predictions compared to standard PINNs. The H-PINN demonstrated a substantial reduction in Mean Absolute Error and Mean Relative Error, and was successfully extended for inverse modeling to determine degradation half-lives from observational data. AI
IMPACT This research advances the application of AI in environmental engineering, potentially leading to more accurate predictions for contaminant transport and material degradation.