Abstract: The balance control of strands tension is an important guarantee for the quality of cable twisting.To solve the problem of closed-loop tension control failure between tension detection of twisting end and the execution end caused by the time-delay effect during tension sensing process,the tension control strategy to cope with the time delay was proposed based on the real-time tension detection data combined with prior knowledge.The time-delay effect of tension data was analyzed theoretically combined with the actual working conditions of cable twisting.The tension signals were decomposed by variational mode decomposition method to obtain modal components with low complexity and high stability that contained multiple original data characteristics.The deep learning method was used to integrate the inherent modal components to realize the prediction and abnormal identification of the twisting tension of strands.Combined with the prior knowledge,a tension balance control system based on prior knowledge of cable twisting was designed.According to the predicted tension of the twisting end,the hysteresis tension of the pay-off wheel at the execution end was adjusted to avoid abnormal tension during the twisting process.The effectiveness of the method was verified by simulation experiments.It provided an effective solution to realize the balance control of tension of cable twisting strand.

Key words: cable twisting, time-delay effect, prior balance control, variational mode decomposition

摘要： 丝股张力均衡控制是线缆捻制质量的重要保障。为解决捻制端张力检测与放线执行端之间的张力感知时滞效应导致闭环张力控制失效的问题,研究了根据张力实时检测数据结合先验知识给出应对时滞的张力控制策略。结合线缆捻制的实际工况,理论分析张力检测数据时滞效应,采用变分模态分解的方法对张力信号进行分解,获得多个包含原始数据特征的低复杂度和较高平稳性的固有模态分量。采用深度学习的方法融合固有模态分量,实现丝股捻制张力的预测与异常识别。结合线缆捻制先验知识,设计基于先验知识的张力均衡控制系统,依据捻制端预测张力调整执行端的放线轮磁滞张紧力,避免出现张力异常。通过仿真实验验证了方法的有效性,为实现线缆捻制丝股的张力均衡控制提供了一种有效的解决方法。

关键词: 线缆捻制, 时间滞后效应, 先验均衡控制, 变分模态分解