Degree Name

Doctor of Philosophy


School of Mechanical, Materials and Mechatronics Engineering - Faculty of Engineering


Thinly cold-rolled steel sheet, following hot-rolling, sporadically present a defect known in industry as ridge-buckle. The defect is only detected in the last stage of production of thin rolled strips, after they have been hot rolled, cold rolled, annealed and tempered. Ridge-buckle consists of sinusoidal waveforms located in one or two bands near the central region of the strip width and in most cases, affecting the full length of the strip. Coils with such a defect have to be discarded or degraded. The aim of this manuscript is to present our contribution towards an elucidation of the origins of ridge-buckle defect by experimental and theoretical approaches combined with analysis of plant data. We explored two of the main factors that in the past, has been regarded as possible origins of ridge-buckle defect: a) the occurrence of microstructural variations along the width of the hot rolled strip, which could lead to different extents of deformation during cold rolling and subsequent operations; and b) the presence of a protuberance (also called ridge) on the hot rolled strip, which will directly affect the shape of the cold rolled strip. The possibility that variations in grain size or crystallographic texture along the width of the strip could contribute to the formation of ridge-buckle defect was explored by studying samples extracted from strips hot-rolled in the industrial process. We determined experimentally and compared quantitatively, the ferrite grain sizes and the crystallographic orientations of grains along the width of the strip. In the strips studied, no evidence was found of any statistically significant variations in these two microstructural variables. Empirical models of recrystallization were used in an attempt to gain a better understanding of the quantitative impact of minor temperature variations (of the order experienced in practice) on the strip strength and hence, its mechanical behaviour within the rolling process. This exploratory exercise confirmed that the strip strength during hot rolling varies very little as a result of relatively small variations in temperature along the width of the strip. Furthermore, the ferrite grain size developed following hot-rolling and cooling on the run-out table, also varies very little due to the temperature variations measured. Therefore, the strength of the hot rolled strip along its width is not affected to the extent that can have a significant influence on the deformation in the subsequent rolling processes. These calculations led us to conclude that ridge in the hot rolled strip is not related to uneven deformation during rolling due to strength variations and hence, ridge buckle defect cannot be attributed to strength differences along the width of the hot-rolled strip. In industrial practice, the occurrence of ridge-buckle in subsequent cold rolling operations has been linked to the presence of ridges in the hot rolled strip. Moreover, a modelling study conducted in parallel with this investigation confirmed this link. Having established that variations in microstructure, texture or strength along the hot-rolled strip width are unlikely causes of the occurrence of ridgebuckle defect we attempted to link ridge occurrence in hot rolled strips and to uneven roll-wear in the finishing mill. By collecting data of 40 consecutive schedules, we were able to determine that ridge formation in hot rolled strips is a common occurrence. Moreover, uneven roll-wear was frequently observed, at least in the period and in the schedules over which our data were collected. A significant finding was the establishment of a strong link between ridge occurrence on the hotrolled strips and uneven roll wear occurring close to the centre of the work-rolls in the finishing-mill. The evidence of uneven wear on the work-rolls in the hot strip mill led our study to an assessment of the possible causes of uneven roll-wear. A thorough review of the available literature in conjunction with the evidence we found of excessive ridge formation in steels containing silicon, led to the hypothesis that the integrity of the tertiary strip scale in the roll bite could be a major contributing factor to uneven roll-wear. We developed a method to study in-situ the formation of oxides on the hot steel substrate, by adapting a High Temperature Microscope with means to introduce air instantaneously onto the surface of the sample. This technique enabled us to study oxidation of the steel surface under conditions closely simulating the oxygen potential pertaining to the rolling temperatures in industrial finishing rolling practice (excluding the complicating effects of water cooling). The in-situ experimentation was complemented with a thorough characterisation of the oxides remaining on the samples after cooling by using relevant metallographic techniques. The experiments were conducted on two types of steel, low carbon steel and silicon containing steel. The results show distinctive differences in the oxidation behaviour of the two types of steel. The low carbon steel form a typical three-layered structure of iron oxides although with very little (almost undetectable) hematite formation. On the other hand, the silicon steels, which apart from silicon also contain aluminium and manganese in small amounts, formed a thick interface layer of complex oxides and intermetallics between the substrate and a layer of magnetite followed by a hematite layer at the gas/oxide interface. It was evident that in silicon steels, the formation of the interface layer interfered with the diffusion of iron species as the final oxide layer was significantly thinner than the oxide layer that formed on low-carbon steel. Another significant finding in the case of low carbon steel was the tendency of the oxide to detach or “spall” from the surface during oxide growth. We were able to observe the detachment of the oxide in-situ with precise account of the time and temperature of occurrence. The relevance of this finding is that if sufficient time elapsed between spallation of the scale and the strip entering the roll-bite, brittle scale would form (due to isolation of the oxide layer from the iron source). The scale then will fracture under the loads applied in the roll-bite, affecting the heat transfer to the rolls and hence, its thermal fatigue resistance. Furthermore, we were able to observe and determine the time to nucleation of new oxides that form on top of already existing oxides on low carbon steel substrate. The time required to nucleate the new oxide is related to the austenite grain size of the steel substrate. Further studies are proposed to elucidate the exact nature of this interesting sequence of events. Finally, early oxidation events were linked to uneven roll-wear due to fracture of the scale in the roll-bite, an increase in the heat transfer rate to the work-roll surface and friction in the roll-bite. An assessment of the possible causes that may affect the integrity of the tertiary scale led us to conclude that whenever uneven changes occur in the cooling patterns of the strip or roll, in descaling efficiency or in the lubrication systems, uneven roll-wear could result. Although a definitive solution to the 40-year old problem of ridge-buckle defect is still to be found, we have made advances in narrowing the possible origins of the defect to specific causes or events. Ridge-buckle is originated mainly by the presence of ridge in the hot-rolled strip; in turn, ridge occurrence is related to uneven wear of the work-rolls in the finishing-mill. Uneven wear is most likely caused by failure of the tertiary scale, due most probably to uneven cooling and/or lubrication patterns during the hot rolling process. In addition to these findings, a novel method of studying scale formation during the early stages of the interaction between oxygen and the steel substrate was developed.