PriMera Scientific Engineering (ISSN: 2834-2550)

Review Article

Volume 4 Issue 4

Inverse Problem Analysis of Phase Fraction Prediction in Aluminum Alloys Using Differentiable Deep Learning Models

Yu Okano*, Takeshi Kaneshita*, Shimpei Takemoto and Yoshishige Okuno

March 20, 2024

DOI : 10.56831/PSEN-04-118


In recent years, there has been an increasing demand for the optimization of alloy properties, driven by the growing complexity of end products and the need to reduce development costs. In general, Thermo-Calc based on the CALPHAD method, which calculates the thermodynamic state of an alloy, is widely used for efficient alloy development. However, a challenge in alloy exploration using Thermo-Calc is the need for specialized computational skills and the significant computational effort required due to the extensive range of calculation conditions for numerous alloys. Consequently, we have developed a deep learning model that rapidly and accurately predicts the temperature-dependent changes in equilibrium phase fractions for 6000 series aluminum alloys (Al-Mg-Si based alloys), which are widely used in industry, using calculations from Thermo-Calc. We developed the architecture of the deep learning model based on the Transformer, which is commonly used in natural language processing tasks. The model is capable of performing calculations more than 100 times faster than ThermoCalc. Furthermore, by leveraging backpropagation of errors in the trained model, we developed a method to estimate the alloy composition for the phase fraction results calculated based on Thermo-Calc.

Keywords: Deep Learning; Inverse Problem; CALPHAD; Transformer


  1. B Sundman, B Jansson and J-O Andersson. “The thermo-calc databank system”. Calphad 9.2 (1985): 153-190.
  2. A Vaswani., et al. "Attention is all you need". in Advances in Neural Information Processing Systems (2017): 5998-6008. Author, F., Author, S., Author, T.: Book title. 2nd edn. Publisher, Location (1999).
  3. Alexey Dosovitskiy., et al. “An image is worth 16x16 words: Transformers for image recognition at scale”. arXiv preprint arXiv:2010.11929 (2020).
  4. Ilya Sutskever, Oriol Vinyals and Quoc V Le. “Sequence to sequence learning with neural networks”. Advances in neural information processing systems 27 (2014).