Aging of Tetrafluoroethylene Propylene (FEPM) Material in High‐Pressure HighTemperature Hydrogen Sulfide Downhole Environment: Theory, Modeling, Experiments, and Material Lifetime Prediction

This paper presents a physics‐driven neural network (PINN) model to capture the constitutive behavior, elongation at break (EAB), modulus, and tensile strength of FEPM material in high‐pressure, high‐temperature (HPHT) hydrogen sulfide (H2S) downhole environment. The model comprises a deep neural network structure and a physics‐based constitutive equation. The PINN model was trained using experimental data from the FEPM material properties before and after the ISO‐23936‐2 immersion tests. It is shown that the PINN model can accurately predict the FEPM material behavior under HPHT sour service conditions. The predictions of our DL model agree well with the experimental data. Predictions were also made via the Arrhenius approach per ISO 23936‐2 based on the same data set, and the results were compared to those from the PINN model. The comparison highlights that the Physics‐Informed Neural Network (PINN) model excels in predicting material longevity across a broad temperature spectrum. This research underscores the effectiveness of PINN models in precisely forecasting material performance in high‐pressure, high‐temperature (HPHT) sour downhole conditions, paving the way for the development of advanced materials.