Author:G.J. Watts- D.C.J. Farrugia- B. Cheong- Z. Husain- M. Zhou I. Gutierrez- D. J. Badiola- J. H. Bianchi- P. Vescovo O. Wiklund- M. Karlberg- M. Schmidtchen- R- Kawalla L. P. Karjalainen- M. C. Somani

Page 3

The key objectives of this project were:

  1. To investigate, by laboratory test simulation and hot rolling trials, the effect on the microstructural evolution of the loading path taken to achieve a given overall cumulative strain.

  2. To identify the mechanisms responsible for differences in constitutive behaviour.

  3. To formulate constitutive models which help provide an understanding of this behaviour. Instances of constitutive models are obtained for a range of steel qualities, including austenitic, ferritic and duplex structures, using experimental data obtained from the laboratory tests, are validated using results from such tests and hot rolling operations to which the models are applied.

    To achieve these objectives the partners have cooperated in supplying materials, in using their experimental testing and rolling equipment to implement the required loading paths, in providing metallurgical services for material characterisation, and in sharing their facilities for development and validation of the constitutive models.

    It has been found, in this work, that comparable mechanical loading paths in the experimental laboratory test simulation trials have given rise to significant differences in rheological and structural behaviour:

  4. An abrupt decrease in strain-rate, in tension tests can give rise to a lag in the flow stress behaviour behind that for a mechanical equation of state, and an even greater lag in the rate of static recrystallisation, and, in the case of ferrite, cells larger than those obtained without the drop in strain rate.

  5. Strain reversal in both tension/compression and torsion tests produces a noticeable Bauschinger effect and a plateau in the flow stress curve. For interpass times greater than 10 s, there is also a change in the static recrystallisation rate depending on the magnitude of the strains applied. For ferrite, the grains are coarser, the cells larger and low angle boundaries more numerous.

  6. A strain reversal coupled with an abrupt decrease in strain rate gives rise, for both torsion and tension/compression, to a Bauschinger effect and plateau more noticeable than without the drop in strain rate. Conversely, if the change in strain rate is an increase, results for torsion exhibit opposite i.e. less pronounced results.

  7. A double hit in the same direction can cause a lag in the recrystallisation curve behind that for the first compression but, if the reductions are sufficiently high, the curve will ultimately be higher than that for a single reduction of the samePage 4total absolute strain. Equivalent triple hits with individually lower reductions may, however inhibit recrystallisation completely. If the two compressions are applied in perpendicular directions, recrystallisation may take longer to complete, although in the short term, the rate may be faster.

    For comparable thermal loading paths in an austenitic steel, differences are insignificant.

    In the laboratory rolling trials, differences have been observed between loads for the high and moderate speed rod rolling schedules and between loads for the monotonic and forward-reverse plate rolling schedules. In the latter application, differences in deformation have also been found. However, unlike experimental tests, effects of loading path on structural evolution, if existent, appear at best to be secondary. Difficulties in detecting any differences have arisen because of the presence of phases such as bainite and wide variations in grain size associated partly with heterogeneity in the initial structure. Equally importantly, the angle between corresponding strain paths is not large, overall, for the simple rolling processes considered.

    To predict the effect of loading path on material behaviour, constitutive models have been formulated by all the partners. In some cases, this has been achieved by incorporating observations from rolling trials or experimental tests, as in the visioplastic model of TU Freiberg used to examine the effect of reversal on deformation in plate rolling and in the empirical model of CEIT for predicting the complete stress-strain curve associated with strain reversal. Other models are more phenomenological in nature, in which the key mechanisms of dislocation dynamics have been incorporated as in the analytical models of the University of Oulu for predicting the stress-strain curve associated with abrupt changes in strain-rate and of CEIT for the simulation of recrystallisation following strain reversal. Generally, these models are numerical, requiring implementation using FE, such as the those used by Corus and MEFOS for predicting the recrystallisation kinetics and evolution of the grain size and dislocation distribution for the H-V rolling trials and the benchmark application viz. the rolling of CMnNb steel plates at TU Freiberg.

    An alternative approach, not including dislocation density, has been followed by CSM in the formulation of a numerical, multilevel model for the simulation of hardening and dynamic recovery, dynamic recrystallisation, and static softening. The model has been set up and validated for a fast recrystallising medium C steel and a non-recrystallising duplex stainless steel using experimental flow stress data for tension/compression and torsion from the University of Oulu and CEIT. The model has been validated for laboratory plate and industrial bar rolling and successfully used in the study of shear reversal.

    For all the models good agreement has generally been obtained between predictions of deformation and load, and experimental measurements and, where the latter can be obtained, of metallurgical structure.

    In conclusion, experimental tests have clearly demonstrated that the flow stress and microstructure can be significantly different for various loading paths used to achieve a given state of equivalent strain, depending on the magnitude of the angle between the paths and of changes in strain rate. Constitutive models have also been formulated and successfully tested and have the potential, in conjunction with appropriate laboratory trials and accurate techniques for material characterisation, for investigating the effects on microstructure, of a wider variation of strain paths in forming processes more complex than those considered hitherto.