摘要
针对传统的PI控制的局限性,提出一种新的控制策略,利用基于神经网络的PI控制器与重复控制器并联构成的控制方法,大大提高了响应速度和补偿精度。该控制器利用BP神经网络技术,根据误差大小对PI控制器参数进行在线实时自适应整定,从而满足最优性的要求。利用重复控制的零稳态误差能力保证了装置的稳态精度。最后通过与传统PI控制器的仿真对比说明所提方法的可行性。
A novel control strategy based on neural network PI algorithm and repetitive control is proposed. Compared with traditional PI controller, this approach has quicker response speed and higher accuracy of compensation. BP neural network technology is adopted to conduct an on-line real-time adaptive setting of PI controller parameter values according to the error value, thus meeting the requirement of an optimal control law. The steady-state accuracy of the device is guaranteed by taking the advantage of the zero steady-state error ability of the repetitive controller. The validity of the proposed method is demonstrated by simulation analysis comparing with traditional PI controller.
出处
《电力系统保护与控制》
EI
CSCD
北大核心
2012年第3期78-84,89,共8页
Power System Protection and Control
基金
国家高新技术863发展计划(2008AA04Z129)
国家自然科学基金(60504010)~~
