Synthetic Tandem Cold Rolling Dataset and Paired Closed-Loop Control Results

Published: 19 November 2026| Version 1 | DOI: 10.17632/rbgr77wbwy.1
Contributors:
Peng-Yuan Fan,

Description

This dataset supports research on Jacobian-supervised multistep prediction and history-assisted model predictive control for coupled thickness–tension regulation in tandem cold rolling. The synthetic dataset comprises 3,000 simulation trajectories generated using a five-stand tandem cold rolling model across diverse material properties and operating conditions, covering MRT4, MRT2.5, and SPCC steel grades. Each trajectory spans 20 seconds with a sampling interval of 0.01 seconds. The dataset also provides paired closed-loop performance results across 48 disturbance pairs, comprising 96 trajectories under mild and severe disturbance conditions. The data are intended to support multistep time-series prediction, learning-based predictive control, disturbance rejection, and control performance evaluation. All shared data are simulation-generated. Raw industrial measurement data are excluded due to data-sharing restrictions.

Files

Institutions

Categories

Computer Science, Engineering

Licence