In collaborazione con: Danieli

Descrizione: In thin-slab continuous casting, the dynamic behavior of the mold meniscus is a critical factor for process stability and product quality. Excessive meniscus fluctuations may affect powder behavior, flow stability inside the mold, and the formation of surface defects. Modern casting machines can be equipped with Ultrasonic Level Detectors (ULDs), which provide real-time, non-contact measurements of mold level and meniscus dynamics. These measurements can be used as feedback signals to support the regulation of electromagnetic braking systems, such as the Multi-Mode Electromagnetic Brake (EMBr). This thesis aims to develop a mathematical and data-driven framework for modeling and controlling meniscus dynamics in thin-slab casting. The first objective is to identify dynamic relationships between ULD signals, process variables, and EMBr operating conditions. Based on this model, optimal or adaptive control strategies may be formulated to determine EMBr setpoints capable of minimizing undesirable meniscus oscillations under varying casting conditions. From a mathematical point of view, the work may involve system identification, time-series analysis, signal processing, dynamical systems modeling, and optimal control techniques. Depending on the available data, machine learning methods may also be investigated to detect abnormal meniscus behavior and support real-time decision-making. The expected outcome is a quantitative framework able to transform real-time process measurements into control-oriented information. This would support improved mold stability, higher product quality, and a more automated management of continuous casting operations.

References:
S. Abouelazayem, I. Glavinić, T. Wondrak, J. Hlava, Adaptive Control of Meniscus Velocity in Continuous Caster based on NARX Neural Network Model, IFAC-PapersOnLine, Volume 52, Issue 29 (2019)
X. Meng, S. Luo, X. Xi, Y. Zhou, W. Wang, and M. Zhu, Characterizing and Controlling Abnormal Periodic Mold Level Fluctuations in a Commercial Slab Continuous Caster Using Big Data, Metallurgical and Materials transactions B, 55B (2024)

Contact: Prof. Dimitri Breda