模糊神经网络在光伏发电MPPT中的应用  

Application of Fuzzy Neural Networks in MPPT of Photovoltaic Power Generating System

作　　者：黄克亚[1] 尤凤翔[1] 李文石[2] 

机构地区：[1]南京铁道职业技术学院,江苏苏州215137 [2]苏州大学电子信息学院,江苏苏州215006

出　　处：《计算机仿真》2012年第8期300-304,共5页Computer Simulation

基　　金：江苏省2009年度高校科研成果产业化推广项目(JHZD09-49)

摘　　要：研究光伏发电系统最大功率点跟踪(MPPT)问题,由于存在随机性,且往往不够准确和充分,容易导致系统无法准确跟踪或稳态剧烈震荡。针对传统的人工总结模糊控制规则有一定难度,提出模糊神经网络控制算法,将神经网络理论和T-S模糊推理方法相结合,选择网格法作为生成算法,混合法作为训练方法,由实测数据自动生成模糊控制规则,并将其嵌入模糊控制器当中去,以实现MPPT控制功能。仿真结果显示,采用该方法生成模糊规则准确实用,系统动态性能和稳态性能均十分优越。实验证明,模糊控制技术与人工神经网络法相结合实现光伏发电MPPT准确高效。In this paper, issues of maximum power point tracking （MPPT） for Photovoltaic power generation sys- tem were discussed. Currently, the fuzzy control rules were generally summed up by the artificial. However, this conclusion is difficult, random, incomplete and inadequate, which may lead to inaccurate tracing or Steady - state turbulent, so fuzzy neural network algorithms were proposed. The fuzzy control rules were adopted using combination of neural network theory and T - S fuzzy reasoning method with measured data, and grid partition was selected as Generation algorithm, and hybrid as Training method. These hales were embedded in fuzzy controller in order to a- chieve the MPPT control. The simulation results show that fuzzy rules generated by this method are accurate and practical, and dynamic performance and steady - state performance are very advantageous. The experiments show the performance of MPPT algorithm becomes more precise and active with the help of fuzzy control and artificial neural network.

关 键 词：最大功率点跟踪 模糊控制 人工神经网络 仿真 

分 类 号：TP391.9[自动化与计算机技术—计算机应用技术]