Computational Materials Science Lab

Research

Computational Design of High Strength Steels

Computational Design of High Strength Steels

PI: Raymundo Arroyave, Ibrahim Karaman

There is a need for a new generation AHSS based on plain carbon and low alloy steels having very low production costs and permitting to retain or further increase the strength achieved with the first generation AHSS, via microstructural and micro-alloying control, at the expense, perhaps, of some of the ductility obtained in the second generation.

 
The main goal of this task is to develop alloying and heat treating guidelines for the design of TRIP-assisted multiphase steels composed of a ferrite matrix with dispersed bainite and relatively high fractions of stabilized retained austenite.


        Computational  Design

(1) Intercritical Annealing Process

Estimating the upper and lower bounds for retained austenite volume fraction.
 (2) Bainite Isothermal Treatment
Determine the effect of alloying and heat treatment on the

(a)phase stability
(b)volume fraction
(c)transformation rate of both Bainiteand retained Austenite.
 (3) Untrafine Grained Alloy

Investigate the stability of retained Austenite under mechanical loading

                                              Theoretical Models

Figure: The phase diagrams for the alloy, Fe-0.32C-1.42Mn-1.56Si. Comparing to the experiments (the “x” points on the right), the thermodynamic and kinetic models predicts the lower and upper bounds of the phase transition

The thermodynamic and kinetic models are implemented. Comparing to the empirical results, the theoretical predictions provides the upper and lower bounds of the carbon enrichments in retained austenite after the BIT treatment.

The optimum heat treated temperatures in two-step heat treatment can be predicted:

 

The red area stands the maximum volume fraction of retained austenite at room temperature which can be approached by the heat treatment. As well as the carbon content and the other phases can be estimated by the theoretical analysis.

                                           Optimization Calculation

For designing/developing a new material, the composition and proper heat treatments are significant. To optimize the micro-structure in TRIP steel, the theoretical models are utilized as the decision maker to evaluate the alloy and Genetic Algorithms (GAs) is coupling to it.


GAs is the computer based algorithm which imitates the natural evolution of the creatures. Using random selection and directional evolving, the extrema in searching domain can be rapidly approached after calculations. Furthermore, the essential “generations” of the calculations are included which avoids the focusing on local extrema values in the domain. 

 
Different Convergence Conditions

In GA, the similarity of the chromosome is an important factor to define the converging of the calculation. In these two picture, two conditions (1% (left) and 5% (right)) similarity are defined in two calculations. It can be seen that in 1% case, the calculation is easily got converged and restart over. So the number of the effective calculations are higher than 5% case.This also means that the 5% case is better to avoid the trapping by the local optima during the calculation.

 
 


Publication:

  1. R. Zhu , S. Li, I. Karaman, R. Arr.oyave, Multi-phase microstructure design of a low-alloy TRIP-assisted steel through a combined computational and experimental methodology, Acta Materialia, 2012.
  2. S. Li, R. Zhu, I. Karaman, R. Arr.oyave, Thermodynamic Analysis of Two-Stage Heat Treatment in TRIP-Steels, Preprint.
  3. S. Li, R. Zhu, I. Karaman, R. Arr.oyave, The kinetic model for simulating the bainitic isothermal transformation, Preprint.
  4. S. Li, R. Zhu, I. Karaman, R. Arr.oyave, A Genetic Algorithm Approach for Designing the microstructure for TRIP Steel, Preprint.