基于干扰观测器的起重机系统神经网络自适应滑模控制  

Neural network adaptive sliding mode control of crane system based on disturbance observer

作　　者：曾兵 陈立庆 饶雄 Zeng Bing;Chen Liqing;Rao Xiong

机构地区：[1]中铁联合国际集装箱有限公司,北京100038 [2]西南交通大学机械工程学院,成都610031

出　　处：《起重运输机械》2025年第5期37-46,共10页Hoisting and Conveying Machinery

基　　金：四川省科技计划项目(2022YFG0241,2023YFG0182)。

摘　　要：文中针对存在负载参数与状态信息未知以及受外界干扰影响的双摆桥式起重机防摇控制问题,提出了基于干扰观测器的神经网络自适应滑模控制策略。首先对建立的起重机四阶欠驱动系统进行分解,单独分析小车和吊钩2个子系统,并为每个子系统分别设计含有非线性函数的新型时变滑模面,基于等效控制原理与双幂次趋近律构建滑模控制律。针对系统中包含负载参数与状态反馈信息的复杂非线性项,利用RBF神经网络结合权值自适应律对其进行逼近。考虑到存在神经网络逼近误差以及外部干扰组成的复合扰动,利用非线性干扰观测器对其进行估计并前馈补偿给控制器,增强系统的抗扰能力。随后通过Lyapunov稳定性理论证明了系统滑模面的稳定性。通过仿真验证了所提控制器优于对比控制器,具有有效的定位、消摆性能,能够对系统参数摄动及多种外界扰动具有鲁棒性。In response to the challenge of anti-sway control for a double-pendulum bridge crane with unknown load parameters and state information,and subject to external disturbances,a neural network adaptive sliding mode control strategy based on a disturbance observer is proposed.Initially,the established fourth-order underactuated system of the crane was decomposed into two distinct subsystems:the trolley and the hook.Each subsystem was analyzed separately.A novel time-varying sliding surface,incorporating a nonlinear function,was designed for each subsystem.Based on the equivalent control principle and a double-power reaching law,a sliding mode control law was constructed.To handle the complex nonlinear terms within the system,including load parameters and state feedback information,a radial basis function(RBF)neural network combined with a weight adaptive law was employed for approximation.To address the complex disturbances arising from both neural network approximation errors and external interferences,a nonlinear disturbance observer was used to estimate the combined disturbances and provides feedforward compensation to the controller,thereby significantly enhancing the system’s anti-disturbance capability.Subsequently,the stability of the system’s sliding surface was rigorously proven using Lyapunov stability theory.Simulation results demonstrate that the proposed controller outperforms the comparative controller.It exhibits superior positioning accuracy and anti-sway performance,and shows remarkable robustness against system parameter perturbations and various external disturbances.

关 键 词：双摆桥式起重机 滑模控制 神经网络 干扰观测器 LYAPUNOV稳定性理论 

分 类 号：TP273[自动化与计算机技术—检测技术与自动化装置]