Exploring structure–property relations in dual phase steels using crystal plasticity and variance based global sensitivity analysis

This study presents a novel framework to quantify the relationships between microstructural features and damage mechanisms in DP800 steel through high-fidelity three-dimensional sRVE simulations with novel damage indicators, which were integrated with variance-based global sensitivity analysis for the calculation of Sobol indices. The developed methodology suggests that the martensite-to-ferrite phase ratio has a stronger impact on damage tolerance than martensite strength, while the elongation of martensite is the dominant parameter for martensite fracture. For the newly introduced phase boundary decohesion indicator, the grain sizes of both phases exhibit the highest influence. A homogenized indicator for overall damage resistance and a trade-off for the two damage mechanisms further revealed the importance of phase morphology, providing insights into additional influencing factors not captured by individual mechanisms. Convergence analyses confirmed that 200–250 datapoints suffice for stable determination of the Sobol indices, confirmed by different surrogate modeling approaches. Radar chart analyses indicated that optimal microstructures for enhanced damage tolerance consist of smaller fractions of strong martensite combined with fine, spheroidal grains in both phases, aligning with established knowledge on DP steels. This approach establishes a validated basis for future optimization of microstructures and loading paths to improve damage tolerance under complex forming conditions. • Development of a novel indicator for phase boundary decohesion in a statistically representative volume element. • Modeling the combined effect of martensite fraction and hardness on the damage of DP steels. • Structure–property quantification based on high-fidelity 3D representative volume element simulations coupled with surrogate modeling and variance based sensitivity analysis.