Multi-objective Optimization of the Process Design in HSLA Steels Using Physically Based Mean Field Modeling

Abstract

A multi-objective optimization approach is presented for the process design of an X70 HSLA steel plate using a physically based mean field (PBMF) model. The PBFM model incorporates an integrated precipitation and recrystallization model to describe the microstructural evolution due to the interaction of strain-induced precipitation of niobium and titanium carbonitrides and static recrystallization of austenite. An integer multi-objective optimization problem is formulated, and the non-dominated sorting genetic algorithm (NSGA-II) is applied on the PBMF model to determine a list of optimal processing routes that satisfy specific process design constraints. Aiming in increasing the strengthening and fracture toughness of the material, a trade-off between the microstructural objectives is considered; average ferrite grain size after accelerated cooling, niobium content in solution and residual strain at the end of multipass hot rolling. Higher residual strains and lower niobium in solution due to higher degree of strain-induced precipitation result in smaller average ferrite grain sizes. An optimal processing route was selected, and the microstructure and precipitation state of the processed material were characterized experimentally revealing a good agreement with model predictions. The proposed approach can contribute to the process design of HSLA and microalloyed steels.