Abstract
Thin steel sheets can be affected by surface defects which are detrimental for the rolling process and spoil the surface appearance. Among other causes, the fracturing of the oxide layer present on the steel surface during the slab hot rolling can end in rafts of hard oxide being embedded into the steel, so called rolled-in scale (RIS). In order to predict this occurrence, a model which identifies the conditions for which the oxide will fracture as a function of steel grade is needed. To identify the conditions causing the oxide fracture, we compared models relative to the oxide properties and to the installation. Once satisfactory results were achieved for experimental data recorded in laboratory conditions, the analysis was performed on data collected in the actual hot rolling mill of TATA Steel IJmuiden. Furthermore, we analysed the results of the first trial of an in-house experiment to directly assess fracturing conditions which can be employed as model calibration for new steels before moving to production.
In conclusion, a model employing the roll forces and the modelled oxide thickness gave promising results for low carbon and low manganese steels. From the factory, the defect classification along the strip length would increase the model accuracy. The lack of an oxide growth model capable to determine the oxide types is limits the analysis of alloyed steel grades. The measure of the strain of fracture needs an improved experimental set-up.
In conclusion, a model employing the roll forces and the modelled oxide thickness gave promising results for low carbon and low manganese steels. From the factory, the defect classification along the strip length would increase the model accuracy. The lack of an oxide growth model capable to determine the oxide types is limits the analysis of alloyed steel grades. The measure of the strain of fracture needs an improved experimental set-up.
Original language | English |
---|---|
Supervisors/Advisors |
|
Award date | 28 Nov 2019 |
Place of Publication | Eindhoven |
Publisher | |
Publication status | Published - 28 Nov 2019 |