Journal of Materials Processing Technology 177 (2006) 486–492 Prediction and avoidance of high temperature damage in long product hot rolling D.C.J. Farrugia Corus RD&T UK, Swinden Technology Centre, Moorgate Road, Rotherham, South Yorkshire S60 3AR, UK Abstract There is currently a drive towards higher contribution steels with improved machinability (in the case of free cutting steels (FCS)) and higher surface quality and consistency. This, together with the potential future implementation of the European legislation (ELVD) to promote lead substitutes in machinable steels (Bi, Te, high S) is leading to the requirement to develop a more thorough understanding of the cause of cracking during hot rolling of bloom and billets of low ductility steels. Physical understanding of the causes and mechanisms of damage initiation and growth (from micro- to macro-scale) at high temperature and relatively high strain rate has not been up to now a major focus of interest, compared to developments in room temperature brittle and ductile fracture, and creep/superplasticity failure. This paper reviews various experimental and modelling approaches (meso- to micro-scale) to develop a better understanding of the influence of thermo-mechanical conditions on damage initiation and growth for as-cast FCS steels. Particular attention is given to the development and use of new/modified mechanical tests. These include double collar and flying saucer axisymmetric tests, U-bending and revised plane strain compression tests using a Gleeble thermo-mechanical simulator to represent the triaxiality, principal stress and strain ratios experienced by the bloom and billet surface during rolling. © 2006 Elsevier B.V. All rights reserved. Keywords: High temperature damage; Free cutting steels; Multiscale modelling; MnS; Inclusions experimental mechanical testing 1. Background There is a need by steel producers to gain a scientific understanding at micro-, meso- and macro-scale of the dam- age mechanisms inducing ductility break-up/defects at both surface and central regions of the deforming stock during hot rolling, for a range steels with inherent low ductility (Fig. 1). Although it is known that intrinsic ductility at high temperature is affected by a range of interacting parameters and processes, from steelmaking (BOS versus EAF route, steel composition, e.g. Mn/S critical index), continuous casting (oscillation marks [1], solidification pattern, cortical zone, MnS inclusions (Fig. 2), etc.), reheating (oxidation, embrittlement) and rolling (cast to wrought structure transition), this paper will only focus on the approach taken to simulate rolling conditions on a Lab- oratory Gleeble thermo-mechanical simulator for a range of low C (<0.1 C) free and leaded-free cutting steels including a range of lead-substitute (Bi, Te) steels. The conditions in terms of key a-dimensional continuum mechanics ratios (triaxiality (STR), stress (PSR) and strain (SPR) principal ratios, etc.) and simple damage criteria integrated over the strain path, for a E-mail address: Didier.Farrugia@corusgroup.com. range of mechanical test specimen geometries are presented and discussed. The conditions tested and simulated are those encountered during cogging/roughing mills, i.e. will replicate the transition from as-cast to wrought structure. The paper also presents a potential framework (still in progress) for addressing the multiscale nature of the problem (except the effect of grain boundaries), i.e. dealing with cavitation, cracking around MnS inclusions or inclusion network (typical inclusion physical scale of 1–2 m at billet corner) to macro-cracking (mm). Care should be taken to differentiate the inclusion from the steel matrix plasticity, which may softened during the initial stage of the process via dynamic/static recrystallisation, in view of the relative small austenite grain size (20–40 m) and the limited grain coarsening (Fig. 3). Indexes of plasticity for type I MnS from [2] have been assumed. Constitutive modelling repre- senting the meso-scale as jointly developed by Birmingham University and Corus [3] but also using the Gurson–Tvergaard [4,5] implementation is also described and referenced [6], together with the need to develop CAFE type modelling [15] using a viscoplastic framework. Redistribution of grain boundary structure with regards to deforming/non-deforming inclusions remains to be addressed and is thought to be an important parameter in the propagation of micro-cracks [7]. 0924-0136/$ – see front matter © 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.jmatprotec.2006.03.236