Mill roll profile control by means of spray cooling
A 2-dimensional model of a mill roll temperature distribution and radial thermal expansion is presented. The emphasis is on selective axial coolant distribution as a method of controlling strip profile. The Fourier equations describing heat conduction are solved using the method of finite differences. The following effects are considered: (i) The temperature distribution within the roll and strip just prior to entry to the roll bite. (ii) The heat generated in the strip due to deformation. (iii) The heat generated by friction between the strip and the roll. (iv) The temperature distribution of the strip and roll af ter each pass. (v) The heat conducted into the roll when in contact with the strip. (vi) The heat removed from the roll by the coolant, the air and the back-up rolls . The model also takes account of the geometry of the roll. This complexity means that different models for the roll and strip temperature distributions are required. The model evaluates individual heat transfer coefficients along the axis of the roll in order to simulate the effectiveness of each spray zone in removing heat from the roll. A simplified method of evaluating roll thermal cambers, derived from the model, is presented. The model shows good agreement between predicted and measured roll thermal cambers.The model is linked to a strip profile prediction model and used to investigate the effects of changing spray patterns and roll bend on profile. It was found that changing spray pattern has a significant effect on strip profile. It was concluded that: (i) Level spray patterns gave the best shape. (ii) Edge sprays sensitivity is important. (ii) over-cooling outside the strip provides good parabolic shape. ' (iv) A change to exit side, spray levels has a significant effect on strip profile. (v) Exit side sprays only has a tendency of rolling out the middle of the slab (i. e. a flat middle). (vi) All level sprays on the exit side of the roll only produce a distorted profile on the strip. It was also found that for any given change in roll thermal camber, there is a corresponding change in strip profile. The two changes can be related by a linear factor. The value of this factor has been investigated and found to be product and mill dependent.