弱耦合星载天线结构振动神经网络预测控制

Vibration control of weak coupling antenna structure in satellite by neural networks predictive method

摘要:

为了研究弱耦合星载天线结构的振动控制,建立了该结构的有限元计算模型;并对结构失谐前后的动力学特性进行了分析;针对失谐前后的结构,分别采用神经网络预测(NNP)控制方法进行了振动控制,并与二次线性最优控制(LQR)方法的振动抑制效果进行了对比.仿真结果表明,弱耦合星载天线结构参数的微小失谐会导致结构振动产生明显的模态局部化;采用NNP控制方法进行结构振动控制的效果明显优于LQR控制方法,且在由失谐导致模型失配时,NNP控制方法对结构振动亦有较好的抑制;在进行此种结构的振动主动控制时必须考虑到结构失谐的影响.

Abstract:

In order to research the vibration control of weak coupling antenna structures in satellite,a dynamics model of this kind of structure is established through the finite element method,by which the dynamic characteristics of the disordered and perfect structures are investigated respectively.Active vibration control of the structure is studied by applying neural networks predictive(NNP) control method and linear quadratic regulator(LQR) control method respectively.Simulation results show that little disorder of the weak coupling antenna structure can induce remarkable mode localization of vibration.The vibration control effects of NNP method are superior to those of LQR method,and NNP method has good performance on the occurrence of model mismatch induced by the structure disorder.The disorder must be considered seriously in active vibration control of weak coupling antenna structures.