Degree Name

Doctor of Philosophy


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


A layer of oxide scale is inevitably formed on the strip surface due to the high temperature of the hot strip rolling process. Experimental result indicates that the oxide scale surface is rough no matter what the original state of the steel surface is. The oxide scale layer and its surface roughness play an important role in the hot strip rolling process. They may affect the friction and heat transfer and consequently affect the rolling force, torque and energy consumption of the rolling process. Surface roughness is also an important index of rolled strip quality.

Hot rolling process and many involved factors have been studied by many researchers. Many models were developed to describe hot strip rolling process and various phenomena in the process. But a model that actually describes the geometrical profile of the rough surface has not been reported. A model that includes oxide scale layer and its surface roughness can better describe the rolling process so that the transformation of the surface roughness can be better understood. The goal of this research is to establish such a model and obtain the required information of the surface roughness transformation and scale layer deformation.

In this thesis, a mathematical model is presented to describe the profile of the surface asperity. The algorithms to generate 3D rough surface or 2D rough surface section through the model are also presented. The model can generate random surface roughness and yet the profiles generated are still relatively simple for the convenience of further FEM analysis. The comparison of generated and scanned surface profiles shows that the model can generate a rough surface that fits very well the actual rough surface of the oxide scale of the strip.

A circular roll model and a flat roll model, which include the oxide scale layer and profiles of the rough surface, have been developed to simulate the hot strip rolling process. With these models, deformation of the oxide scale layer and the transformation of the surface roughness in the rolling process can be calculated. Other important rolling parameters such as pressure distribution and real area of contact ratio can also be obtained through such models. Comparing the parameters obtained through simulation with those from the experiments, the models developed are effective in calculating these parameters. A developed thermo-mechanical coupling FEM analysis has also been implemented with such models.

Many factors may affect the deformation of oxide scale layer and the surface roughness transformation in the hot strip rolling process. The effects of these factors such as 􀂾 Flow stress of oxide scale and steel materials 􀂾 Initial roughness 􀂾 Reduction 􀂾 Work roll surface roughness 􀂾 Scale layer thickness 􀂾 Lubrication 􀂾 Voids inside the scale layer 􀂾 Coefficients of friction on the interfaces of tool-scale and scale-steel 􀂾 Wavelength of the asperity

on the scale layer deformation and the surface roughness transformation have been evaluated through the developed models. Simulation results compare closely to the experimental results, which are obtained from the experiments in Gleeble 3000 and experimental rolling mill Hille 100, in scale surface roughness change and other parameters for the hot strip rolling process.

The oxide scale layer may crack before or inside the roll bite. The development of the pre-existed openings in the oxide scale layer is simulated. The effect of the ratio of the initial opening width over scale layer thickness, frictional coefficient on the tool-scale and scale-steel interfaces, profile roughness of the interfaces and the lubrication on the propagation of the crack in the rolling process is explored. A model of the crack generation from a weak position in the oxide scale layer has also been developed and the final crack width is in agreement with experimental results.

After careful calculation and comparing with the experimental results, a series of conclusions is drawn and further work for the research in this area is recommended.

02Whole.pdf (7979 kB)