Machine learning-based optimization of crystal plasticity model parameters

Thumbnail Image

Files

URL

Journal Title

Journal ISSN

Volume Title

Insinööritieteiden korkeakoulu | Bachelor's thesis
Electronic archive copy is available locally at the Harald Herlin Learning Centre. The staff of Aalto University has access to the electronic bachelor's theses by logging into Aaltodoc with their personal Aalto user ID. Read more about the availability of the bachelor's theses.
Unless otherwise stated, all rights belong to the author. You may download, display and print this publication for Your own personal use. Commercial use is prohibited.

Authors

Date

2025-05-09

Department

Major/Subject

Computational Engineering

Mcode

ENG3082

Degree programme

Aalto Bachelor’s Programme in Science and Technology

Language

en

Pages

30

Series

Abstract

This thesis develops and implements a machine learning-based approach for optimizing crystal plasticity (CP) model parameters in a complex multiphase Q&P1000 steel. Traditional parameter calibration is time-consuming due to the large number of parameters and computational costs of CP simulations. By utilizing Long Short-Term Memory (LSTM) neural networks trained on lattice stress data from approximately 2400 CP simulations, this study demonstrates an efficient calibration workflow that predicts constitutive parameters across four distinct phases (ferrite, tempered martensite, new martensite, and retained austenite). The model achieves high predictive accuracy for yield parameters (R² scores up to 0.96) while exhibiting varying performance for hardening parameters. Validation against synchrotron X-ray diffraction experimental data shows excellent agreement for lattice stresses in BCC phases with R2 scores of nearly 0.9, while challenges remain for accurately modeling retained austenite behavior due to phase transformation complexities. The results highlight the potential of machine learning for accelerating CP model calibration while identifying specific areas for improvement, particularly in handling transformation-induced plasticity effects. This approach advances the field of integrated computational materials engineering by enabling rapid, microstructure-informed parameter identification for complex advanced high-strength steels.

Description

Supervisor

St-Pierre, Luc

Thesis advisor

Juan, Rongfei

Keywords

crystal plasticity, machine learning, optimization, QP1000

Other note

Citation

Permanent link to this item

https://urn.fi/URN:NBN:fi:aalto-202507295630

Collections

[kand] Insinööritieteiden korkeakoulu / ENG