Modelling the microstructural evolution during hot rolling
G.R. Gómez And T. Pérez
Dto. de Materiales y Corrosión, CINI-FUDETEC, 2804 Campana, Buenos Aires, Argentina
sidggz@siderca.com
Abstract — A metallurgical model that describes the microstructural evolution of C-Mn steels in the hot strip mill, and that predicts the yield strength and the ultimate tensile strength based on the steel chemistry and the processing conditions is presented. The model comprises the microstructural evolution during hot deformation in the austenitic range (taking into account the effects recrystallization and grain growth), and the austenite decomposition during cooling (formation of ferrite, perlite and/or bainite). A comparison between calculated and measured yield strength and ultimate tensile strength values for several steels and processing conditions is also included.
Keywords — hot rolling, recrystallization, grain growth, phase transformation.
I. INTRODUCTION
Improving the hot strip mill process is mandatory for
the steel industry owing to the increasingly severe specifications
being imposed by end users. Within this objective
the development of a metallurgical model that
links the operating variables in the mill with microstructure
and mechanical properties of the final products
becomes of great importance. This model could provide
useful information for process control and optimization,
reducing the need of costly on-line experimentation.
The aim of this work is to present a metallurgical
model that could describe the microstructural evolution
of C-Mn steels during hot rolling, and could predict the
final mechanical properties based on the steel chemistry
and the processing conditions.
During hot deformation the average dislocation density of the material increases several orders of magnitude.
Two metallurgical processes become active to
reduce the dislocation excess: recovery and recrystallization.
In austenite, the later is the most important restoration
process. Depending on temperature, strain and
strain rate the recrystallization could begin during deformation
(dynamic recrystallization) or in the interpass
time (static recrystallization). In order to describe the
microstructural evolution it is needed a description of
the recrystallization progress and the resulting austenite
grain size after each reduction step in the rolling mill. In
addition further modifications of the grain size could
take place if the recrystallization after deformation is
complete and is followed by grain growth.
After hot deformation the austenite transforms to
ferrite, perlite and/or bainite. The transformation kinetic
and final phase distribution depend strongly on the austenite
microstructure (grain size and accumulated
strain), and on the cooling and coiling conditions. Due
to the important metallurgical differences between the
several phases that may appear, it is needed an accurate
description of the final microstructure in order to evaluate
the mechanical properties of the material.
II. EXPERIMENTAL PROCEDURE
Stress relaxation tests (Liu and Jonas, 1988) performed
with a Gleeble 3500 thermo-mechanical simulator have
been used to measure the kinetic of recovery and recrystallization
for different processing conditions (deformation
temperature, strain and strain rate). In order to
study the recrystallized grain size and the kinetic of
grain growth, samples were deformed in compression
and quenched after different holding times. The boundaries
of the prior austenite grains were revealed using
standard metallographic methods.
Dilatometric tests were also performed in the Gleeble
machine to study the kinetic of austenite decomposition
into ferrite, perlite and/or bainite.
Several low carbon-manganese steels were analyzed,
the chemistries were within the following ranges
[C] = 0.09-0.17 wt%, [Mn] = 0.8-1.2 wt% and [Si] =
0.1-0.3 wt%.
III. HOT ROLLING MODEL
In the present model the average grain size and the accumulated
strain have been used to characterize the microstructure
of the steel in the austenitic range. For the
calculation of the microstructural evolution during hot
deformation we have considered the effects of recrystallization
and grain growth.
If the accumulated strain in a single reduction is
lower than the critical value (ec) for the initiation of dynamic
recrystallization, static recrystallization proceed
after deformation (interpass time). The kinetic of static
recrystallization was modelled using an Avrami type
equation as proposed by Sellars (1980):
![]() |
(1) |
where X is the volume fraction of the material recrystallized, t0.5 is the time to 50% recrystallization and n is an empirical parameter. Several expressions have been proposed to describe the t0.5 dependence with austenite grain size (d0), strain (e), strain rate (de/dt) and holding temperature (T) (see Devadas et al., 1991). In the present model the functional form proposed by Hodgson and Gibbs (1992) was adopted:
![]() |
(2) |
The parameters of this equation (At, p, q, r, Qrex) were fitted using experimental data from stress-relaxation tests. Some of the results obtained with this technique are presented in Fig. 1, where the recrystallization start and finish times for a C-Mn steel are plotted as a function of temperature.
There are many empirical expressions that attempt to
describe the austenite grain size after static recrystallization
(see Devadas et al., 1991). In our case the predictions
obtained with IRSID equations (Perdrix, 1987)
presented the best agreement with experimental data.
The critical strain value to produce dynamic recrystallization
was determined using the following expression
(Sellars, 1980):
![]() |
(3) |
where Z is the Zener-Hollomon parameter
![]() |
(4) |
that depends
on temperature and strain rate. We assume that when
dynamic recrystallization takes place, it is followed by
metadynamic recrystallization. In the present model the
equations proposed by Hodgson and Gibbs (1992) to
describe the kinetic of metadynamic recrystallization
and to evaluate the size of the recrystallized grains were
used.
The grain growth kinetic was modelled using a well
known isothermal law (Sellars, 1980):
where drec is the fully recrystallized grain size, t is the time after complete recrystallization, and m, k and Qg are empirical parameters that were adjusted as was described in Section II.
IV. PHASE TRANSFORMATION
To describe the austenite decomposition into proeutectoid
ferrite, perlite and bainite an empirical model based
on the Miettinen work was used (Miettinen, 1996). The
start temperature for each transformation was calculated
with empirical expressions that take into account the
chemical composition of the steel (C, Mn, Si), the
cooling rate, and the austenite grain size. The effect of
the accumulated strain during hot rolling was considered
using an effective austenite grain size (Suehiro et
al., 1987).
The progress of the proeutectoid ferrite, perlite and
bainite reactions was described using Avrami type
equations:
![]() |
(5) |
where Xf is the volume fraction occupied by f phase at time t (f: ferrite, perlite or bainite), and Xeq is the thermodynamic equilibrium fraction for proeutectoid ferrite or the maximum available volume fraction in the case of perlite and bainite. The following dependence of k with the temperature was adopted:
![]() |
(6) |
The empirical parameters (n, Pi) that appear in Eqns. 5 and 6 were fitted for each reaction using data obtained from dilatometric tests. Experimental results and the comparison with the model predictions are shown in Fig.2.
The final ferrite grain size was calculated using a semi-empirical expression proposed by Suehiro et al. (1987), with parameters adjusted by comparison with experimental results. The interlamellar spacing of perlite was estimated as l = 5 leq, where leq corresponds to the equilibrium interlamellar spacing (static interphase between austenite and perlite, see Christian, 1975).
V. MECHANICAL PROPERTIES
The metallurgical model has been used to predict the
mechanical properties of hot rolled C-Mn steels. Due to
the steel chemistries covered in this study and the applied
cooling rates, the final microstructures were mixtures
of proeutectoid ferrite and perlite. The yield
strength (YS) and ultimate tensile strength (UTS) were
calculated using the set of equations developed by
Gladman et al. (1972). These equations take into account
the effects of the steel chemistry, the volume
fraction of ferrite and perlite, the ferrite grain size, and
the interlamellar perlitic spacing.
The model input data (strains, strain rates, deformation
temperatures and interpass times) were calculated from
the hot rolling mill set up (reduction schedule, strip velocities,
average temperatures, work roll diameters and
rotational speeds). The coiling temperature and the
cooling bed operation parameters were used to estimate
the cooling rate during transformation. In Figs. 3 and 4 a
comparison between the calculated and measures values
of YS and UTS is shown for several C-Mn steels. It can
be seen in these graphs that there is a good agreement
between measured and theoretical results.
There are some discrepancies between theoretical and experimental values of YS and UTS that can be ascribed to carbon and manganese contents out of the experimentally adjusted chemical range reported in Section II. The model parameters were adjusted for carbon and manganese compositions higher that 0.09 wt% and 0.8 wt%, respectively. The materials that exhibit higher deviations in the YS values have C concentration between 0.04 wt% and 0.05 wt% and Mn concentrations between 0.4 wt% and 0.6 wt%. In these cases some differences in the kinetics of recrystallization and grain growth that were not taken into account in the model may appear and, in addition, there could be changes in the rate of the austenite decomposition that affect the final ferrite grain size.
VI. SUMMARY
A metallurgical model that comprises the prediction of the microstructural evolution during hot rolling (austenitic range) and the subsequent phase transformation was developed for C-Mn steels. This model allows the prediction of some of the final product mechanical properties (YS and UTS). The calculated values of YS and UTS present a good agreement with experimental results. We have observed some discrepances that we have ascribed to steels chemistries outside the range covered when the empirical parameters of the model were adjusted.
REFERENCES
1. Christian, J.W., Theory of transformations in metals andalloys, 2nd ed. Part I, Pergamon Press, Oxford,
United kingdom (1975).
2. Devadas, C., I.V. Samarasekera and E.B. Hawbolt, "The
thermal and metallurgical state of steel strip during
hot rolling: Part III. Microstructural Evolution", Metall. Trans. A22, 335-349 (1991).
3. Gladman, T., I.D. McIvor and F.B. Pickering, "Some
aspects of the structure-property relationships in
high-carbon ferrite-perlite steels", J. Iron and Steel
Inst. 210, 916-930 (1972).
4. Hodgson, P.H. and R.K. Gibbs, "A mathematical model
to predict the mechanical properties of hot-rolled
C-Mn and microalloyed steels", Iron Steel Inst.
Jpn. Int. 32, 1329-1338 (1992).
5. Liu, W.J. and J.J. Jonas, Metall. Trans. A19, 1403
(1988).
6. Miettinen, J., "Simple semiempirical model for prediction
of austenite decomposition and related heat
release during cooling of low alloyed steels", Ironmaking and Steelmaking 23, 346-356 (1996).
7. Perdrix, C., "Characteristics of plastic deformation of
metals during hot working", Report CECA No.
7210 EA/311, The institute de Rechearches de la
Siderurgie Francaise (IRSID), Saint Germainen-
Laye, France (1987).
8. Sellars, C.M., "The physical metallurgy of hot working". Proc. Int. Conf. On Hot Working and Forming
Processes, Eds. C.M. Sellars and C.J. Davies,
The Metal Society of London, 3-15 (1980).
9. Suehiro, M., K. Sato, Y. Tsukano, H. Yada, T. Senuma
and Y. Matsumura, "Computer modeling of microstructural
change and strength of low carbon steel
in hot strip rolling", Trans. Iron and Steel Inst.
Jpn. 27, 439-445 (1987).
Received: May 15, 2001.
Accepted for publication: December 15, 2001.
Recommended by Subject Editor E. Dvorkin.