基于神经网络的双馈电机的分数阶滑模控制  

Fractional Sliding Mode Control Based on Neural Network for Doubly Fed Generator

作　　者：管萍[1] 李军 GUAN Ping;LI Jun(School of Automation,Beijing Information Science and Technology University,Beijing 100192,China)

机构地区：[1]北京信息科技大学自动化学院

出　　处：《电气传动》2019年第8期75-79,共5页Electric Drive

基　　金：国家自然科学基金资助项目(61573230)

摘　　要：针对不平衡电网电压条件下双馈电机运行性能差的问题,将神经网络算法与分数阶滑模结合起来应用于双馈电机直接功率控制中。利用分数阶微积分的遗传衰减特性与神经网络的自适应性,削弱传统滑模控制中的抖震。用神经网络逼近电机模型的不确定项,推导了分数阶滑模控制律。仿真结果表明,所提出的控制系统实现了消除有功功率波动和转矩波动的目标,有效抑制了传统滑模控制中存在的抖震。Considering that doubly fed induction generator works poorly under the condition of unbalanced grid voltage,neural network and fractional sliding mode control were combined,then it was applied to the direct power control system of doubly fed induction generator. By using the genetic attenuation characteristics of fractional calculus and the adaptive ability of neural network,the chattering in the traditional sliding mode control was weaken. The neural network was used to approximate the uncertainty of the induction generator model,and the fractional order sliding mode control law was derived. The simulation results show that the stator active power fluctuation and electromagnetic torque fluctuation are eliminated in the proposed control system and the chattering in traditional sliding mode control is effectively restrained.

关 键 词：双馈电机 不平衡电网 神经网络 分数阶 

分 类 号：TM46[电气工程—电器]