基于改進的單神經(jīng)元三相并網(wǎng)逆變器控制策略研究
滿燦
(南京理工大學,江蘇 南京 210094)
摘 要:為改善LCL型三相并網(wǎng)逆變器控制器無法根據(jù)外界條件校正參數(shù)的缺點,提出了一種基于改進的單神經(jīng)元三相并網(wǎng)逆變器控制策略。該控制策略基于并網(wǎng)電流、電容電流反饋的雙電流環(huán)控制策略,將并網(wǎng)電流外環(huán)用單神經(jīng)元PI自適應控制器代替,根據(jù)有監(jiān)督的Hebb學習規(guī)則對自身權值系數(shù)進行更新,利用辨識器校正PI參數(shù),實現(xiàn)系統(tǒng)動態(tài)調節(jié)KP、KI值,仿真結果表明,該控制策略增強了系統(tǒng)魯棒性、穩(wěn)定性。
關鍵詞:LCL 型濾波器;單神經(jīng)元;自適應控制
中圖分類號:TM464 文獻標識碼:A 文章編號:1007-3175(2016)12-0023-04
Research on Control Strategy of Three-Phase Grid Connected
Inverter Based on Improved Single Neuron
MAN Can
(Nanjing University of Science and Technology, Nanjing 210094, China)
Abstract: In order to improve the disadvantage that the LCL type of three-phase grid connected inverter isn’t able to revise the parameters according to the external conditions, this paper proposed a new control strategy based on an improved single neuron three-phase grid connected inverter. The grid connected current outer ring was replaced by the single neuron proportion integration (PI) self-adaptive controller. According to the supervised Hebb learning rule, one's own weight coefficients were updated. The identifier was used to revise the PI parameters, so as to dynamically adjust the KP and KI values. The simulation results show that this control strategy enhances the robustness and stability of the system.
Key words: LCL type filter; single neuron; self-adaptive control
參考文獻
[1] 由世俊,婁承芝,華君,等. 建筑物用光伏集成系統(tǒng)在中國應用的前景[J]. 太陽能學報,2000,21(4):434-438.
[2] 曾正,趙榮祥,湯勝清,等. 可再生能源分散接入用先進并網(wǎng)逆變器研究綜述[J]. 中國電機工程學報,2013,33(24):1-12.
[3] 鄭昕昕,肖嵐,王長寶,等. 三相變流器LCL濾波器參數(shù)優(yōu)化新方法[J]. 中國電機工程學報,2013,33(12):55-63.
[4] MALINOWSKI M,BERNET S.A simple voltage sensorless active damping strategy for threephase PWM converters with an LCL filter[J]. IEEE Transactions on Industry Electronics,2008,55(4):1876-1880.
[5] 徐志英,許愛國,謝少軍. 采用LCL濾波器的并網(wǎng)逆變器雙閉環(huán)入網(wǎng)電流控制技術[J]. 中國電機工程學報,2009,29(27):36-41.
[6] 劉飛,查曉明,段善旭. 三相并網(wǎng)逆變器LCL濾波器的參數(shù)設計與研究[J]. 電工技術學報,2010,25(3):110-116.