摘要
针对直驱系统中各种非线性干扰直接作用下的永磁直线电机控制性能恶化的问题,提出一种深度模糊滑模-自抗扰控制方法。首先,通过非线性扩张状态观测器估计系统不确定扰动,在滑动面设计中引入跟踪误差的积分项,结合饱和函数sat(s)与位移误差的幂函数设计趋近律,从而改进滑模-自抗扰控制方法;其次,为避免设计过程中的主观影响,进一步提升直线电机在复杂工况下的适应能力,基于深度神经网络训练模糊规则,进而调节滑模控制的关键参数。采用所提方法在直驱泵性能测试平台进行了实验研究,结果表明:所提深度模糊滑模-自抗扰控制方法有效提高了直线电机控制精度、响应速度与鲁棒性,直线电机的阶跃响应时间相对于传统滑模控制方法提升了23.87%;正弦目标跟踪的ITAE指标改善是传统滑模控制方法的4.7%、是改进滑模-自抗扰控制方法的13.2%;在传感器白噪声干扰下,正弦目标跟踪的最大跟踪误差相对于改进滑模-自抗扰控制方法改善了一个数量级。
A depth fuzzy sliding-mode active disturbance rejection control(DFSM-ADRC) method is proposed to solve the problem that the control performance of permanent magnet linear motors(PMLM) deteriorates under the direct action of various nonlinear disturbances. Firstly, the system uncertainty disturbance is estimated by a nonlinear extended state observer. The integration of tracking error is applied to the design of the sliding surface, and the saturation function sat(s) and the power function of displacement error are introduced into the approach law, so that the improved sliding mode active disturbance rejection control(ISM-ADRC) method is proposed. Secondly, to avoid the subjective influence and further improve the adaptability of PMLM under complex conditions, fuzzy rules are trained by deep neural networks, and the key parameters of sliding mode control are adjusted. The experimental research has been carried out through the performance test platform of the direct drive pump. The results show that the proposed DFSM-ADRC method effectively improves the control accuracy, response speed and robustness of the controller. The proposed DFSM-ADRC method improves the step response time by 23.87%, compared with the conventional sliding mode control(SMC);improves the ITAE of sinusoidal target tracking by 4.7% and 13.2%, compared with the SMC method and the ISM-ADRC method respectively;and improves the maximum tracking error of sinusoidal target tracking under sensor white noise by one order of magnitude, compared with the ISM-ADRC method.
作者
谭草
鲁应涛
葛文庆
李波
陆佳瑜
TAN Cao;LU Yingtao;GE Wenqing;LI Bo;LU Jiayu(School of Transportation and Vehicle Engineering,Shandong University of Technology,Zibo,Shandong 255000,China)
出处
《西安交通大学学报》
EI
CAS
CSCD
北大核心
2023年第1期185-194,共10页
Journal of Xi'an Jiaotong University
基金
国家自然科学基金资助项目(51905319,51875326)
山东省青年科技人才托举工程项目(SDAST2021qt20)。
关键词
永磁直线电机
直驱泵
深度神经网络
滑模控制
自抗扰控制
permanent magnet linear motor
direct drive pump
deep neural network
sliding mode control
active disturbance rejection control
作者简介
谭草(1991-),男,副教授;通信作者:葛文庆,男,教授,博士生导师。
