As the demand for thin and ultra-thin gage for cold-rolled sheet metals increases, cold rolling manufacturers and sheet metal producers must utilize technological advancements to stay competitive globally. To manufacture a competitive rolling mill, the mill designer must identify an appropriate mill configuration, roll sizes, and flatness control systems, etc. The design must be robust enough to accommodate broad product envelopes containing strips of different widths, gages, and materials. At the same time, the mill “sizing” must be competitive by avoiding over- or under-design. These two objectives, designing a robust mill and remaining competitive, tend to conflict, which establishes the need for optimal mill design based on rigorous multidisciplinary analysis.
Conventionally, mill design is done using a ‘sequential’ design approach relying heavily on experience. The major limitation of sequential mill design is that it can lead to over-design or under-design since it segregates design tasks for the coupled components, making it difficult to consider many combinations of the design space. The result is a ‘feasible’ rather than ‘optimum feasible’ mill design. This work applies nonlinear programming to optimize the basic design of a 4-high reversing cold rolling mill, subject to static strength, fatigue, and strip geometric quality constraints. The work offers an improved multi-disciplinary design approach based on stress and fatigue analysis of rolls, and an efficient strip profile/flatness analysis to identify suitable design limits for roll bending devices.
Decoupling Predictive Control of Strip Flatness and Thickness of Tandem Cold Rolling Mills Based on Convolutional Neural Network