非对称基神经网络跟踪光伏最大功率点的方法

发布时间：2018-03-11 10:48

本文选题：模糊技术　切入点：非对称基　出处：《西安电子科技大学学报》2017年05期 　论文类型：期刊论文

【摘要】：为了使光伏发电系统时刻工作在最大功率点处,构建了非对称基神经网络跟踪光伏最大功率点的方法,给出了该方法的具体实现步骤.依据光伏发电因素对发电效率的影响程度不同,构建了模糊因素隶属函数,计算出影响因素的模糊权值,并将该权值融入到非对称基神经网络结构的构建中.通过固定基宽的径向基函数方法、传统的径向基函数方法以及文中方法,并采用4种数量的样本训练网络,通过网络训练时间及标准差进行对比,可得采用180个样本训练网络的精度最高,且文中方法获得网络的精度高于其他方法至少1个数量级以上.使用这种神经网络时刻识别光伏系统的工作参数,能使光伏系统通过滑动变阻器在任一时刻均能让内外电阻完全匹配,从而保证该系统时刻工作在最大功率点处.
[Abstract]:In order to make the photovoltaic system work at the maximum power point at all times, an asymmetric basis neural network is constructed to track the photovoltaic maximum power point. The realization steps of the method are given. According to the different influence degree of photovoltaic power generation factors on generation efficiency, the membership function of fuzzy factors is constructed, and the fuzzy weights of the influencing factors are calculated. The weight value is incorporated into the construction of asymmetric basis neural network structure. Through the fixed basis width radial basis function method, the traditional radial basis function method and the method in this paper, four kinds of samples are used to train the network. By comparing network training time and standard deviation, 180 samples can be used to train the network with the highest accuracy. The accuracy of the proposed method is at least one order of magnitude higher than that of other methods. The neural network is used to identify the working parameters of photovoltaic system at all times. The photovoltaic system can make the internal and external resistor match exactly at any time by sliding rheostat, so that the system can work at the maximum power point at all times.
【作者单位】：信阳师范学院计算机与信息技术学院;
【基金】：国家自然科学基金资助项目(61402393) 河南省高等学校重点科研资助项目(16A535001) 河南省教师教育课程改革研究重点资助项目(2017-JSJYZD-022,2017-JSJYYB-221) 河南省高等教育教学改革资助项目(2017SJGLX260)
【分类号】：TM615;TP183

【相似文献】