滑模动态递归模糊神经网络船电推进复合控制
Hybrid control based on sliding mode--dynamic recursive fuzzy neural network for marine electrical propulsion
摘要点击 1638  全文点击 1822  投稿时间:2010-05-28  修订日期:2010-08-21
查看全文  查看/发表评论  下载PDF阅读器
DOI编号  10.7641/j.issn.1000-8152.2011.5.CCTA100630
  2011,28(5):625-630
中文关键词  复合控制  鲁棒滑模  动态递归模糊神经网络  船舶电力推进  吊舱推进
英文关键词  hybrid control  robust sliding mode  dynamic recursive fuzzy neural network  marine electrical propulsion  podded propulsion
基金项目  国家“863”计划资助项目(2006AA09Z308); 国家自然科学基金资助项目(50979058).
作者单位E-mail
张桂臣 上海交通大学 海洋工程国家实验室 zhanggc2004@163.com 
马捷 上海交通大学 海洋工程国家实验室  
中文摘要
      提出了船舶电力吊舱推进系统的复合控制策略, 以消除吊舱推进的过冲现象并获得快速平滑的动态响应. 复合控制由鲁棒滑模控制和动态递归模糊神经网络控制组成, 鲁棒滑模控制利用死区非线性和误差边界厚度法, 克服系统的不确定与外界扰动, 具有在线自学习算法的动态递归模糊神经网络控制促使系统的跟踪误差趋近于0. 建立了基于SIMOTION的半实物仿真Siemens-Schottel推进器系统, 仿真与实验结果表明, 复合控制具有暂态快速和稳态平滑的动态响应, 提高了吊舱推进系统的鲁棒性和运动精度.
英文摘要
      We propose a hybrid control(HC) strategy for the marine electrical podded propulsion system to eliminate the overshoot and obtain a fast and smooth dynamic response for the podded propulsion. HC consists of a robust sliding mode control(SMC) and a dynamic recursive fuzzy neural network control(DRFNNC). SMC uses the dead-zone nonlinearity and error band method to tackle uncertainties and external disturbances; DRFNNC which has online self-learning algorithm forces the tracking error to approach zero. We build the hardware-in-loop simulation system of Siemens-Schottel-Propulsor(SSP) based on SIMOTION; the simulation and experimental results show that HC provides a fast and smooth dynamic response in both transient state and steady state, and improves the robustness and motion precision of the SSP system.