模糊神经网络在光伏发电M中的应用.docVIP

欢迎引用：[1] 黄克亚, 尤凤翔, 李文石. 模糊在光伏发电MPPT中的应用模糊在光伏发电MPPT中的应用 摘要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生成模糊控制规则。动态和稳态性能。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。关键词模糊控制人工神经网络最大功率点跟踪 仿真 中图分类号：T　 文献标识码：AFuzzy Neural Networks in MPPT of Photovoltaic Power Generating System HUANG Ke-ya1 , YOU Feng-xiang1, LI Wen-shi2 (1. Nanjing Institute of Railway Technology, Suzhou Jiangsu, 215137; 2. Department of Microelectronics, Soochow University, Suzhou Jiangsu, 215006) ABSTRACT：In this paper, issues of maximum power point tracking (MPPT) for Photovoltaic power generation system were discussed. Currently, the fuzzy control rules were generally summed up by the artificial. However, this conclusion is difficult, random, incomplete and inadequate. These will 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, hybrid as Training method. These rules were embedded in fuzzy controller, in order to achieve the MPPT control. The simulation results show fuzzy rules generated by this method were accurate and practical, and dynamic performance, steady-state performance were very advantageous. The experiments show the performance of MPPT algorithm becomes much precise and active with the help of fuzzy control and artificial neural network. KEYWORDS：Maximum power point tracking; Fuzzy control; Artificial neural network; Simulation 1 引言 目前已知MPPT算法主要有固定电压法、扰动观察法、电导增量法、模糊控制法和神经网络法等。固定电压法简单且容易实现，但跟踪效果较差，现在已较少使用。扰动观察法和电导增量法同为MPPT经典算法，但扰动观察法抗干扰能力较差，容易发生误判现象，电导增量法对硬件要求较高，需要高精度的A/D转换器；两种算法跟踪步长设置大小都会影响系统的动态和稳态性能，且较难找到平衡点；这些特性都限制了两种算法的应用。 模糊控制是一种仿人思维的控制技术，它不依赖于被控过程的数学模型。神经网络对环境变化有极强的自习能力，在建模方面具有黑箱学习模式特点。在学习完成后，从输入、输出数据得出的关系却无法用人们易于接受的方式表示出来。[1]。 论文尝试将模糊理论表达知识能力和神经网络的自学能力结合起来，Matlab神经网络工具箱，由实测数据自动产生模糊控制规则，并将其嵌入到光伏MPPT模糊控制过程当中 光伏阵列MPPT原理 光伏组件的输出存在着功率最大点，在特定的温度和光照条件下，组件能否工作在最大功率点取决于组件所负载