MODELING IMPACT TOUGHNESS OF INDUSTRIAL HOT ROLLED HSLA STEELS

A quantitative model has been developed to predict the temperature dependence of impact toughness (KCV) for hot rolled plane carbon and HSLA steels with a wide range of chemical composition (wt%): С(0.04-0.21), Mn(0.16-1.96), Si(0.01-0.93), Cr(0.01-0.97), Ni(0.01-0.51), Cu(0.02-0.45), Mo(0.001-0.50), Nb(0.001-0.056), V(0.002-0.09), Ti(0.001-0.08), N(0.003-0.009), S(0.001-0.032), P(0.005-0.025). The transition temperature at 50 % of fibrous fracture ( FATT 50 ) is determined in terms of chemical composition and calculated final microstructure parameters. Impact energies at lower ( KCV LS ) and upper ( KCV US ) shelves are derived from predicted values of tensile stress and relative elongation. All physical parameters of the model are related to the industrial hot rolling conditions by means of the integral computer model STAN 2000 previously formulated for the rolling mill 2000 of PJSC Severstal. Empirical coefficients are fitted to the data base on impact toughness for 230 plates of 32 steel grades in the temperature range of -60°С to +20 °С. The modelling results comply well with experiments.


INTRODUCTION
Apart from characteristics of strength (yield stress and ultimate tensile stress) and plasticity (relative elongation), the temperature-dependent impact energy is a crucial property that determines reliability of steel constructions exploited under dynamic loads and low temperatures. Accordingly, considerable attention is paid to experimental studies of the impact toughness of high strength steels with various microstructures and chemical compositions and development of the corresponding predictive models [1][2][3][4][5][6][7].
Cold resistance of the high strength steels is often specified by the 50% Fibrous Fracture Appearance Transition Temperature (FATT50). One of the first models to predict this characteristic for ferritic-pearlitic steels in terms of ferrite grain size, thickness of cementite plates at grain boundaries and other structural features was developed in [2]. Next [4,5], the obtained dependences have been improved and empirical expressions for impact energies at lower (KCVLS) and upper (KCVUS) shelves were proposed for several steels with more complex microstructures.
In this paper, original empirical expressions for calculating FATT50, as well as for KCVLS and KCVUS, are proposed depending on chemical composition, final microstructure parameters and mechanical properties of industrial hot-rolled steels with ferrite-pearlite and bainite microstructures. Using these physical parameters, a model is composed for calculating KCV as a function of temperature.

INVESTIGATED STEELS AND DATA USED TO CALIBRTATE THE MODEL
To develop the present model, a data base has been employed that covers 230 plates of 32 steels with a wide range of chemical composition ( Microstructure parameters involved in simulations, as well as mechanical properties (tensile strength and relative elongation), were calculated using previously developed computer model for hot rolling STAN 2000 [8]. All computations correspond to protocols of industrial modes of rolling and accelerated cooling. According to respective results, 158 of investigated plates had ferritic-pearlitic structures (volume fraction of pearlite up to 25 %) and 72 plates had bainitic structures (mostly granular bainite). Related mechanical properties vary in wide ranges: 250-780 MPa of yield stress, 360-880 MPa of ultimate tensile stress and 10-40 % of relative elongation.

DESCRIPTION OF THE MODEL, MODELING RESULTS AND DISCUSSION
The temperature dependence of impact toughness has been treated according to [3][4][5] with the mixture rule: Following [4,5], the ductile fraction is expressed by: for such effects, we will make use of FATT as a characteristic most sensitive to them. To further facilitate the modeling, specific approximations will be applied to the upper and lower shelves of KCV.
To determine FATT (°С) at various chemical compositions and microstructures, contributions of different physical mechanisms [1,2,4,5] are allowed for: where successive terms respectively correspond to the contributions of:  w . = (7) Note that the effect of S is much stronger than that of P. Probably, this is due to an implicit allowance of MnS inclusions also reducing the cold resistance. Contributions of various precipitates are expressed as follows: -contents of free microalloying elements, calculated with allowance for their interaction with carbide or carbonitride particles TS  (MPa),  (%) -ultimate stress and relative elongation at room temperature Here, the product of ultimate stress and relative elongation at room temperature reflects the virtual plastic work. As to the lower shelf, its invariant impact energy is employed: (20) Although Eq. (20) ignores actual KCVLS variations, the latter do not notably affect predicted FATT50. Besides, from the practical viewpoint, these variations as such do not matter insofar as the considered ductile-to-brittle transition temperature commonly limits admitted working conditions. At the same time the employed KCVLS seemingly overestimate the true (least) impact toughness of the lower shelf; to determine it when necessary, tests should be performed at lower temperatures (say, T < 80 °С).

CONCLUSION
The model has been developed for evaluating impact toughness of industrial hot rolled steels with ferriticpearlitic and bainitic microstructures in dependence of chemical composition, combination of microstructural parameters and mechanical properties. All these characteristics were derived from the hot rolling conditions by means of previously developed integral computer model of hot rolling STAN 2000. The present modelling results are in good agreement with the experimental data. Average value of the relative error of calculations over the entire set of considered steel strips is 14.6 %.