Abstract :
[en] Thinner and harder steel strips, which are in great demand in the production of lighter car bodies, challenge cold rolling mills like never before. The objective of this thesis is therefore to accurately model friction in lubricated cold rolling to ultimately minimize friction while preventing skidding between the rolls and the strip by flexible lubrication, i.e. by adjusting the lubrication conditions in real-time depending on the current rolling conditions.
Accordingly, the most comprehensive experimental data of lubricated cold rolling are analyzed to identify the underlying physical mechanisms that have to be included in the rolling model. Furthermore, these data are post-processed to test this model.
By means of the previous data, the most powerful model of lubricated cold rolling available so far is completely rederived, documented, and improved while its computer implementation is entirely refactored. This model, which is called Metalub, is a 2D cold rolling model in the mixed lubrication regime with a thermo-elastoviscoplastic description of the strip, a thermo-piezoviscous description of the lubricant, non-circular roll flattening and lubricant starvation. In particular, strain rate hardening and thermal softening are added to the model in addition to improvements regarding its robustness.
Metalub is then tested based on the previous experimental data to evaluate its predictive capabilities and shortcomings. After the first systematic calibration of its numerical parameters, this model allowed to improve predictions of earlier research by more physically consistent hypotheses. Nevertheless, the model is still limited by unavailable material parameters, the simplicity of its thermal model as well as manual adjustments of the lubricant film thickness and the boundary coefficient of friction, if starvation or micro-plasto-hydrodynamic/static (MPH) lubrication (permeation of the lubricant from microscopic surface pockets into the solid/solid contact zone between asperities) occur, respectively.
Thus, the first coupling procedure between Metalub and the finite element (FE) solver Metafor is developed to replace analytical asperity flattening equations by FE simulations of lubricated asperity flattening and to ultimately include MPH lubrication in the model.
Finally, lubricated asperity flattening is for the first time simulated by the Lagrangian meshless particle method "smoothed particle hydrodynamics" (SPH) to eliminate mesh-related weaknesses of FE asperity flattening models.
