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

作　　者：张桂臣[1] 马捷[1] 

机构地区：[1]上海交通大学海洋工程国家实验室,上海200030

出　　处：《控制理论与应用》2011年第5期625-630,共6页Control Theory & Applications

基　　金：国家"863"计划资助项目(2006AA09Z308);国家自然科学基金资助项目(50979058)

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.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 nonlin-earity 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.

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 

分 类 号：U665.13[交通运输工程—船舶及航道工程]