基于改进SSA结合模糊RBF神经网络的悬臂梁振动主动控制

Active Vibration Control of Cantilever Beam Based on Improved SSA Combined with Fuzzy RBF Neural Network

随着航空航天事业的发展,为了节省燃料,同时提高航天器速度,航天器采用更轻的材料来减少质量。然而,此举也引入了柔性振动,灵活的振动增加了姿态控制的时间,导致姿态精度控制不尽如人意。因此,有效抑制柔性振动以实现高精度姿态控制非常重要。论文以柔性压电悬臂梁作被控对象,并利用压电薄膜(Polyvinylidene Fluoride,PVDF)作传感器和致动器,分析其振动的控制问题。基于PID和模糊理论的局限性,结合模糊控制器能模仿专家经验和径向基神经网络(Radial Basis Function Network,RBFNN)善于学习的优点,设计了模糊径向基(Fuzzy Radial Basis Function,FRBF)神经网络控制器来抑制悬臂梁的振动,并采用混沌映射的种群初始化策略、疯狂算子的领导者位置更新策略、精英保留及动态惯性权重的追随者位置更新策略改进的樽海鞘群算法(Salp Swarm Algorithm,SSA)来优化模糊神经网络权值。将改进后的控制方法在Matlab软件环境下进行了数值仿真,仿真结果表明,应用改进的模糊径向基神经网络控制器可以有效地提升主动控制的振动效果。

With the development of aerospace industry,in order to save fuel,while improving the speed of spacecraft,spacecraft use lighter materials to reduce mass.However,it also introduces flexible vibration,which increases the time of attitude control and leads to unsatisfactory attitude precision control.Therefore,it is very important to suppress flexible vibration effectively to achieve high precision attitude control.In this paper,the vibration control of flexible piezoelectric cantilever beam is analyzed by using piezoelectric film(Polyvinylidene fluoride(PVDF))as sensor and actuator.Based on the limitations of PID and fuzzy theory,combined with the advantages that Fuzzy controller can imitate expert experience and RBFNN is good at learning,a fuzzy radial basis function is designed.FRBF neural network controller is used to suppress the vibration of cantilever beam,and the improved salp swarm algorithm is adopted by using chaotic mapping population initialization strategy,crazy operator leader position update strategy,elite retention and dynamic inertial weight follower position update strategy to optimize the weights of fuzzy neural networks.The simulation results show that the improved fuzzy RBF neural network controller can effectively improve the vibration effect of active control.