Skip to content

A library based on Tensorflow for constructing neural-operator surrogate models for dynamical systems using the DeepONet architecture

Notifications You must be signed in to change notification settings

sandialabs/romnet

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

12 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

ROMNet: Neural Network Based Surrogates for Reduced-Order Model Dynamics


Executing ROMNet:

  1. In $WORKSPACE_PATH, create a ROMNET folder

  2. Inside $WORKSPACE_PATH/ROMNet/, clone the ROMNet repository and rename it "romnet"

    Note: $WORKSPACE_PATH ├── ... ├── ROMNET │ └── romnet ├── app ├── database ├── ...

  3. From $WORKSPACE_PATH/ROMNet/romnet/, install the code (i.e., $ python3 setup.py install)

  4. From $WORKSPACE_PATH/ROMNet/romnet/app/, launch the code (i.e., $ python3 RomNet.py path-to-input-folder) (e.g. fpython3 RomNet.py ../input/MassSpringDamper/DeepONet/)


Test Cases:

Please, refer to the presentations in $WORKSPACE_PATH/ROMNet/romnet/docs/test_cases for info and running instructions

It is highly recommended to run the set test cases in the following order:

- 1. Mass-Spring-Damper System

- 2. Translating Hyperbolic Function

- 3. 0D Isobaric Reactor 

Implemented NN Algorithms:

  • Deterministic Neural Networks

  • Probabilistic Neural Networks

    • Monte Carlo Dropout

    • Bayes by Backprop

About

A library based on Tensorflow for constructing neural-operator surrogate models for dynamical systems using the DeepONet architecture

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published