基于神经网络逼近的磁浮列车动态悬浮控制

为了有效提高磁浮列车悬浮系统在负载扰动和轨道不平顺扰动下的动态特性,提出了一种基于Lyapunov稳定性分析的径向基神经网络逼近算法使悬浮间隙能够在有界范围内达到最优。首先,以悬浮负载为受控对象建立系统垂向动力学方程和电压控制方程,以此构造能够表征系统非线性的状态空间方程。其次,确定径向基函数(radial basis function, RBF)神经网络基本结构,根据悬浮间隙误差约束条件和控制电流构造输入输出,并以此设计控制律保证所输出悬浮间隙能够在多种扰动的综合作用下持续稳定;再次,基于Lyapunov稳定性第二判据证明系统闭环稳定,能够在误差整定过程中使得间隙误差收敛于无穷小。最后,通过与目前应用较为广泛的比例-积分-微分(proportion-integral-derivative, PID)控制算法进行仿真对比,在非线性负载力和不平顺扰动下分析验证所提出控制算法的有效性。结果表明:所提控制算法比PID控制算法具有更好的鲁棒性。

Research on dynamic levitation control of Maglev Train Based on neural network approximation

In order to effectively improve the dynamic characteristics of maglev train levitation system under load disturbance and track irregularity disturbance, a radial basis function neural network approximation algorithm based on Lyapunov stability analysis is proposed in this paper, so that the levitation gap can be optimized in a bounded range. Firstly, the vertical dynamic equation and voltage control equation were established by taking the suspended load as the controlled object, and the state space equations were constructed to represent the nonlinearity of the system. Secondly, the basic structure of RBF(Radial basis function) neural network was determined, and the input and output were constructed according to the suspension gap error constraints and control current. The control law was designed to ensure that the output suspension gap can be continuously stable under the combined action of various disturbances; Thirdly, based on the second Lyapunov stability criterion, the closed-loop stability of the system was proved, which can make the gap error converge to infinitesimal in the error tuning process. Finally, the effectiveness of the proposed control algorithm was verified by simulation comparison with PID(proportion- integral-derivative) control algorithm, which is widely used at present. The results show that the proposed control algorithm has better robustness than PID control algorithm