基于RBF神经网络单神经元PID控制在低压直流变换中的应用研究  

Application Research on Low Voltage DC Conversion Based on Single Neuron PID Control of RBF Neural Network

作　　者：刘学文[1] 任兴贵 徐定杰[2] LIU Xuewen;REN Xinggui;XU Dingjie(Guangzhou Huaxia Vocational College,Guangzhou 510640,Guangdong Province,China;Harbin Institute of Technology,Harbin 150000,Heilongjiang Province,China)

机构地区：[1]广州华夏职业学院,广东广州510640 [2]哈尔滨工业大学,黑龙江哈尔滨150000

出　　处：《电力与能源》2019年第3期280-282,287,共4页Power & Energy

基　　金：国家自然科学基金(61573117)

摘　　要：太阳能和风能发电系统的非线性特征非常明显,并且二次端直流负载的接入与断开的随机性也比较大,容易引起微型电网的波动。为了解决低压直流微型电网的波动特性,提出了基于RBF神经网络在线辨识的单神经元PID控制的直流变换方法,该方法结合了普通PID控制结构简单和神经网络非线性系统自适应性强的特性,解决了二次直流负载端稳定的直流电源供电问题。With the widespread application of clean solar energy and wind energy, the island-style micro DC power grids are used in remote areas and isolated islands. In these places, the irregular characteristics of the atmosphere often result in the obvious nonlinear characteristics of the solar and wind power generation system,and the large randomness of the access and disconnection of the secondary-side DC load,which is easy to cause the fluctuations of the micro grid. In order not to slove the fluctuation charateristics of low voltage Dcmicrogrid. the single-neuron PID control DC conversion based on RBF neural network online identification is proposed in this paper combined with the advantages of both the simple structure of PID control and the strong adaptability of neural network to nonlinear systems, conducive to solving the problem of secondary DC load stability.

关 键 词：RBF神经网络 单神经元 PID控制 直流变换 

分 类 号：TM359.4[电气工程—电机]