一种带有前馈补偿的L1神经网络风机最大功率点跟踪控制方法  

An L1 Neural Network Maximum Power Point Tracking Control Method for Wind Turbines with Feedforward Compensation

作　　者：屈繁 胡慧[1] 徐睿婷 彭龙 颜渐德[1] 袁君奇 陈柏志 QU Fan;HU Hui;XU Ruiting;PENG Long;YAN Jiande;YUAN Junqi;CHEN Bozhi(College of Electrical and Information Engineering,Hunan Institute of Engineering,Xiangtan 411104,China;Xiangtan Hengxin Industrial Co.,Ltd.,Xiangtan 411101,China;Xiangtan Traction Locomotive Factory Co.,Ltd.,Xiangtan 411100,China)

机构地区：[1]湖南工程学院电气与信息工程学院,湘潭411104 [2]湖南华菱湘潭钢铁有限公司,湘潭411101 [3]湘潭牵引机车厂有限公司,湘潭411100

出　　处：《湖南工程学院学报（自然科学版）》2023年第2期14-21,共8页Journal of Hunan Institute of Engineering(Natural Science Edition)

基　　金：湖南省自然科学基金资助项目(2020JJ6021、2021JJ50115);湖南省教育厅重点项目(21A0451、20A116).

摘　　要：针对风机在最大功率点跟踪(MPPT)控制中由于自身存在非线性和未知干扰导致跟踪困难的问题,提出一种带有前馈控制的L1神经网络(L1NN)自适应最大功率点跟踪控制方法.其中的前馈控制器用于补偿叶片转矩带来的干扰项的影响,通过神经网络权值的在线学习,对系统的非线性因素进行逼近,在反馈回路中设有低通滤波器.当满足L1的增益条件时,系统具有最终有界性,通过提高自适应增益来满足所需的瞬态性能.最后在不同风速条件下,将本文所提方法与无前馈补偿方法以及传统模型自适应方法进行仿真实验对比,实验结果表明:本文方法在提高系统响应速度和效率方面均有明显优势.Aiming at the difficulty of fan tracking in MPPT due to nonlinearity and unknown interference,an adaptive MPPT control method based on L1 neural network(L1NN)with feedforward control is proposed in this paper.A feedforward controller is used to compensate for the disturbance of blade torque item.Through the online learning of the neural network weights,the nonlinear factors of the system are approaching,and a low-pass filter is designed in the feedback loop.When the gain conditions for L1 are met,system has the ultimate boundedness to improve the transient performance of adaptive gain to meet the required.Finally,the proposed method is simulated with no feedforward compensation method and the traditional model adaptive method under different wind speed conditions.The experimental results show that the proposed method has obvious advantages in improving the response speed and efficiency of the system.

关 键 词：风力发电 神经网络 前馈补偿 L1自适应控制 

分 类 号：TP273[自动化与计算机技术—检测技术与自动化装置] TM315[自动化与计算机技术—控制科学与工程]