Robust Algorithm for Calculating the Alignment of Guide Rolls in Slab Continuous Casting Machines

To ensure the product quality of a steel slab continuous casting machine, the mechanical alignment of the guide rolls must be monitored and corrected regularly. Misaligned guide rolls cause stress and strain in the partially solidified steel strand, leading to internal cracks and other quality issues. Current methods of alignment measurement are either not suited for regular maintenance or provide only indirect alignment information in the form of angle measurements. This paper presents three new algorithms that convert the available angle measurements into the absolute position of each guide roll, which is equivalent to the mechanical alignment. The algorithms are based on geometry and trigonometry or the gradient descent optimization algorithm. Under near ideal conditions, all algorithms yield very accurate position results. However, when tested and evaluated under various conditions, their susceptibility to real-world disturbances is revealed. Here, only the optimization-based algorithm reaches the desired accuracy. Under the influence of randomly distributed angle measurement errors with an amplitude of 0.01°, it is able to determine 90% of roll positions within 0.1 mm of their actual position.