Sign in Subscribe
Fatigue (Materials)

Data-Driven Approach to Modeling Creep-Fatigue Behavior Using Neural ODEs

Strategic Insights

— 2 min read

Research conducted by Argonne National Laboratory has presented a data-driven machine learning approach to modeling one-dimensional stress-strain behavior under cyclic loading, utilizing experimental data from the nickel-based Alloy 617. The approach employs uniaxial creep-fatigue test data acquired under various loading histories and compares two distinct neural network-based ODE models. Financial support for this research came from the United States Department of Energy (DOE). The study has demonstrated the potential of neural network-based ODE models to depict complex creep-fatigue behavior, eliminating the necessity for experts to define a specific, material-focused model form.

Key Takeaways:

  • The study introduced a data-driven machine learning approach for modeling one-dimensional stress-strain behavior under cyclic loading, utilizing experimental data from the nickel-based Alloy 617.
  • The approach employs uniaxial creep-fatigue test data acquired under various loading histories and compares two distinct neural network-based ODE models.
  • The research was conducted with financial support from the United States Department of Energy (DOE) and has been peer-reviewed.
  • The study demonstrated that the neural network-based ODE models precisely capture the experimental creep-fatigue mechanical behavior, exceeding the standard Chaboche model's accuracy.
  • An interpretable model derived from the black-box neural ODE model through symbolic regression achieved accuracy comparable to the Chaboche model, enhancing its interpretability.
  • The approach eliminates the necessity for experts to define a specific, material-focused model form.
  • The research was published in the journal Optimization and Engineering in 2025.
  • The study was conducted by researchers at Argonne National Laboratory, who obtained a quote from the research stating that the results highlight the potential of neural network-based ODE models to depict complex creep-fatigue behavior.

Statistics:

  • The research was conducted with financial support from the United States Department of Energy (DOE).
  • The study utilized experimental data from the nickel-based Alloy 617.
  • The approach employed uniaxial creep-fatigue test data acquired under various loading histories.
  • The neural network-based ODE models precisely captured the experimental creep-fatigue mechanical behavior with an accuracy exceeding the standard Chaboche model.

Sources:

  • A Data-driven Method for Modeling Creep-fatigue Stress-strain Behavior Using Neural Odes. Optimization and Engineering, 2025.
  • Argonne National Laboratory, 9700 Cass Ave, Lemont, IL 60439, United States. Hao Deng, Appl Mat Div, United States.
  • United States Department of Energy (DOE).
  • NewsRx. Study Results from Argonne National Laboratory Provide New Insights into Information Technology (A Data-driven Method for Modeling Creep-fatigue Stress-strain Behavior Using Neural Odes). Information Technology Newsweekly. November 4, 2025; p 941.

Read more