SPAC: S parse sensor p lacement-based a daptive c ontrol for high precision fuselage assembly
SPAC: S parse sensor p lacement-based a daptive c ontrol for high precision fuselage assembly
复制标题
SPAC：基于 Sparse 传感器放置的自适应控制，用于高精度机身组装
DOI：
10.1080/24725854.2022.2116133
复制
发表时间：
2022
影响因子：
2.6
通讯作者：
Shi, Jianjun
中科院分区：
文献类型：
--
作者：
Mou, Shancong;Biehler, Michael;Yue, Xiaowei;Hunt, Jeffrey H.;Shi, Jianjun
最佳形状控制在机身装配过程中很重要。对于存在初始形状偏差的机身，为了实现高精度装配，需要进行形状调整。最先进的方法通过使用致动器来实现这一目标，致动器的力来自基于所设计的机身的机械特性的模型。这有一个很大的局限性：它们没有考虑由于单个进入机身的形状偏差引起的机械性能变化而引起的模型失配。模型失配会导致控制性能恶化。为了提高性能，需要根据机身形状调整的在线反馈信息更新形状控制模型。然而，由于机身表面的大尺寸，在实践中获得高精度的内联测量是昂贵的或甚至是不可行的。为了解决这些问题，本文提出了一种基于稀疏传感器布局的自适应控制方法。在该方法中，基于响应信号的稀疏传感器测量来更新模型。理论上量化了较小模型失配下的重建性能。其性能已被评估的基础上的真实的数据的半到半机身装配过程中，所提出的方法提高了控制性能与可接受的传感努力。
Optimal shape control is important in fuselage assembly processes. To achieve high precision assembly, shape adjustment is necessary for fuselages with initial shape deviations. The state-of-the-art methods accomplish this goal by using actuators whose forces are derived from a model based on the mechanical properties of the designed fuselage. This has a significant limitation: they do not consider the model mismatch due to mechanical property changes induced by the shape deviation of an individual incoming fuselage. The model mismatch will result in control performance deterioration. To improve the performance, the shape control model needs to be updated based on the online feedback information from the fuselage shape adjustment. However, due to the large size of the fuselage surface, highly accurate inline measurements are expensive or even infeasible to obtain in practice. To resolve those issues, this article proposes a Sparse sensor Placement-based Adaptive Control methodology. In this method, the model is updated based on the sparse sensor measurement of the response signal. The reconstruction performance under a minor model mismatch is quantified theoretically. Its performance has been evaluated based on real data of a half-to-half fuselage assembly process, and the proposed method improves the control performance with acceptable sensing effort.