上海交通大学学报 ›› 2022, Vol. 56 ›› Issue (2): 165-172.doi: 10.16183/j.cnki.jsjtu.2020.424

• • 上一篇    下一篇

六自由度波浪补偿平台的神经网络自适应反馈线性化控制

丁明1, 孟帅1(), 王书恒1, 夏玺2   

  1. 1. 上海交通大学 海洋工程国家重点实验室,上海 200240
    2. 中国船舶重工集团公司第七一一研究所,上海 201203
  • 收稿日期:2020-12-14 出版日期:2022-02-28 发布日期:2022-03-03
  • 通讯作者: 孟帅 E-mail:mengshuai001@sjtu.edu.cn
  • 作者简介:丁明(1997-),男,山东省青岛市人,硕士生,从事波浪补偿平台研究.
  • 基金资助:
    国家自然科学基金资助项目(51879161)

Neural-Network-Based Adaptive Feedback Linearization Control for 6-DOF Wave Compensation Platform

DING Ming1, MENG Shuai1(), WANG Shuheng1, XIA Xi2   

  1. 1.State Key Laboratory of Marine Engineering, Shanghai Jiao Tong University, Shanghai 200240, China
    2.Shanghai Marine Diesel Engine Research Institute, Shanghai 201203, China
  • Received:2020-12-14 Online:2022-02-28 Published:2022-03-03
  • Contact: MENG Shuai E-mail:mengshuai001@sjtu.edu.cn

摘要:

六自由度波浪补偿平台所采用的大长径比非对称液压系统在深海区需完成大跨度、高速度的波浪补偿任务,这为控制系统的控制精度和抗干扰能力带来严峻的挑战.引入径向基神经网络(RBFNN)辨识,提出一种自适应反馈线性化控制策略.首先,建立六自由度波浪补偿平台非对称液压系统的非线性模型.然后,基于RBFNN辨识利用反馈线性化设计自适应控制器.最后,利用MATLAB/Simulink开展五级海浪(90°遭遇角恶劣工况)作用下和外力干扰下的仿真分析.结果表明:相比于经典比例系数-积分系数-微分系数(PID)和滑模控制,新型控制器控制精度和抗干扰能力明显提高,更适合用于复杂海况下六自由度波浪补偿平台的控制,且具有很好的跟踪效果和较强的稳健性,可为深海区六自由度波浪补偿平台控制系统设计提供参考.

关键词: 波浪补偿平台, 径向基神经网络, 液压系统, 反馈线性化, 自适应控制

Abstract:

Ocean resource exploration expands into deep and ultra-deep waters, which has posed great challenges to the 6-DOF parallel platform that requires to finish the long-span and high-velocity wave compensation task with high precision and anti-interference ability. The control strategy employed in the asymmetric hydraulic system of large aspect ratio requires more careful considerations when operating in the harsh and severe environment. An adaptive feedback linearization control strategy was proposed by employing the radial basis function neural network (RBFNN) for identification. First, a nonlinear model of the asymmetric hydraulic system was established. Then, an adaptive controller was designed based on RBFNN and feedback linearization. Finally, simulations were performed by using MATLAB/Simulink under the five-stage wave environment at a 90° wave encounter angle and under the external interference condition. The result shows that this method has a good traceability and robustness compared to classic PID and sliding mode control methods, which is more suitable in control of the wave compensation platform in complex sea conditions. The new controller can significantly increase the compensation accuracy and anti-interference ability, and provide a workbench for the 6-DOF parallel platform operation in deep waters.

Key words: wave compensation platform, radial basis function neural network, hydraulic system, feedback linearization, adaptive control

中图分类号: