WEB Predicting damage in dual phase steels by means of simulation and machine learningWednesday (23.09.2020) 10:10 - 10:25 M: Modelling and Simulation 2 Part of:
Dual Phase (DP) steels are an important family of steel grades used by the automotive industry. They consist of soft ferritic grains, which are mainly responsible for ductility, and hard martensitic zones, which give these steels their strength. The high mechanical contrast among ferrite and martensite and the different types of morphologies inherited from the specific processing conditions makes DP steels susceptible for internal damage evolution. Ideally, microstructures of DP comprise complex shaped ferrite grains surrounded by smaller martensite particles. While this microstructure gives rise to desired properties such as high ultimate tensile strength (GPa range), low initial yield stress, high early-stage strain hardening (a feature which is very important for crash parts), and macroscopically homogeneous plastic flow, it also introduces many potential damage initiation sites. More specifically, the high mechanical contrast between ferrite and martensite leads to a very inhomogeneous stress—strain distribution which leads to various damage mechanisms such as martensite cracking, decohesion between grains and phases, and void formation. Understanding and predicting damage in DP steels therefore requires therefore studying which micromechanical features trigger these damage mechanisms. To tackle this challenging task, we evaluate results from spectral method-based crystal plasticity simulations by means of data science and machine learning.