Detail

Ostwald Ripening Dataset for "Accelerating phase-field simulation of coupled microstructural evolution using autoencoder-based recurrent neural networks"

Gesch, Aidan H.; Hu, Chongze

Year

2025

Source Name

34c73a10-2b22-404c-9c6f-652569720a28

License

Creative Commons Attribution 4.0

Contacts

ahgesch@crimson.ua.edu hucz@ua.edu

DOI

10.18126/ccqv-t437 View on Datacite
Collection of phase field simulations of the Ostwald Ripening phenomenon, generated for use in "Accelerating phase-field simulation of coupled microstructural evolution using autoencoder-based recurrent neural networks" by Gesch, A. and Hu, C. Each file is a Numpy array corresponding to one simulation frame. The first section each file name is the simulation number, from 0000_out to 0199_out. The second section of each file name is the tracked phase: composition, ostwald 1, ostwald 2, ostwald 3, or ostwald 4. The last digits of each file name is the simulation timestep, from 0 to 30,000,000 with an interval of 30,000 for a total of 101 timesteps included per simulation per phase. Each array contains only 0s and 1s, with 0 corresponding to out of phase and 1 corresponding to in phase. At each timestep, the sum of all four individual phases will equal the compositional phase. The simulation parameters used were randomized but checked to ensure mictrostructures generated. A dump with these parameters is available in excel format from the authors.