Welding is an important part in the construction process of large cruise ships. Seam tracking is essential in the welding processes to ensure the welding accuracy. To reduce seam tracking error in the process of steel sheets welding for large cruise ships, a digital-twin-based predictive compensation control strategy is proposed. Firstly, based on an analysis of ideal characteristics for welding processes, a digital-twin-based architecture of predictive compensation control for seam tracking is introduced. Then, the basic logic of predictive compensation control strategy for seam tracking is explained. In addition, a recursive grey model based on Taguchi method and sparrow search algorithm (RGMbTM-SSA) is developed to improve prediction accuracy. Finally, the accuracy of RGMbTM-SSA and the feasibility of the proposed digital-twin-based predictive compensation control strategy are verified through actual cases of steel sheets welding production for large cruise ships. The results show that in the prediction of seam tracking errors, RGMbTM-SSA performs better than some common grey models and machine learning algorithms. In the developed digital twin system for seam tracking, the comprehensive performance of predictive compensation control strategy achieves the best when the advance time is set to 2 s. The strategy proposed in this study can reduce the average over-offset number per seam by more than 60 %. Simultaneously, approximately 20 % of average offsets can be reduced. Digital-twin-based predictive compensation control strategy enriches the methods of seam tracking, and it can significantly improve the quality of welding seams.

中文翻译：

---

基于数字孪生的大型游轮钢板焊接焊缝跟踪预测补偿控制策略

焊接是大型游轮建造过程中的重要环节。焊缝跟踪在焊接过程中至关重要，以确保焊接精度。为了减少大型游轮钢板焊接过程中的焊缝跟踪误差，提出了一种基于数字孪生的预测补偿控制策略。首先，基于对焊接工艺理想特性的分析，引入了基于数字孪生的焊缝跟踪预测补偿控制架构。然后，解释了焊缝跟踪预测补偿控制策略的基本逻辑。此外，还开发了基于田口方法和麻雀搜索算法的递归灰色模型（RGMbTM-SSA）以提高预测精度。最后，通过大型游轮钢板焊接生产的实际案例验证了RGMbTM-SSA的准确性以及所提出的基于数字孪生的预测补偿控制策略的可行性。结果表明，在焊缝跟踪误差的预测中，RGMbTM-SSA比一些常见的灰色模型和机器学习算法表现更好。在所开发的焊缝跟踪数字孪生系统中，当提前时间设置为2 s时，预测补偿控制策略的综合性能达到最佳。本研究提出的策略可以将每条接缝的平均过度偏移数量减少 60% 以上。同时，可以减少大约 20% 的平均偏移。基于数字孪生的预测补偿控制策略丰富了焊缝跟踪的方法，可以显着提高焊缝质量。