非对称基神经网络跟踪光伏最大功率点的方法

doi:10.3969/j.issn.1001-2400.2017.05.022

Abstract

Abstract:

To ensure that the photovoltaic power generation system always works at its maximum power point, a method for photovoltaic maximum power point tracking by the neural network based on asymmetric basis is proposed, and its concrete implementation steps are given. Fuzzy factor membership functions are built according to the influences of photovoltaic power generation factors on the power generation efficiency, and the fuzzy weights of the influencing factors are calculated, with these weights infused into the building of the neural network based on fuzzy asymmetric basis. The network is trained by using methods of fixed basis width RBF, traditional RBF and the method proposed in this paper with four kinds of quantities of samples, and the comparison in terms of the network training time and the standard deviation indicates that the accuracy of the network with 180 samples is the highest, at least an order of magnitude higher than other that of methods. By determining the working parameters of the photovoltaic system in real time by using this neural network, the photovoltaic system can make the internal and external resistances completely match at every moment through the slide rheostat, thus ensuring that the system always works at the maximum power point.

Key words: fuzzy technology, asymmetric basis, neural network (NN), photovoltaic, maximum power point

LIU Daohua;HU Xiuyun;LIU Jiangfeng;GUAN Feng. Method of tracing the photovoltaic maximum power point based on NN with asymmetric basis[J].Journal of Xidian University, 2017, 44(5): 127-133.

References

［1］ CHANG F K, HUANG Y C, JENG J S, et al. Band Offset of Vanadium-doped Molybdenum Oxide Hole Transport Layer in Organic Photovoltaics［J］. Solid-State Electronics, 2016, 122(4): 18-22.
［2］ LOPEZ-GUEDE J M, RAMOS-HERNANZ J A, ZULUETA E, et al. Systematic Modeling of Photovoltaic Modules Based on Artificial Neural Networks［J］. International Journal of Hydrogen Energy, 2016, 41(29): 12672-12687.
［3］ BANA S, SAINI R P. A Mathematical Modeling Framework to Evaluate the Performance of Single Diode and Double Diode Based SPV Systems［J］. Energy Reports, 2016, 2: 171-187.
［4］ PYRGOU A, KYLILI A, FOKAIDES P A. The Future of the Feed-in Tariff(FiT) Scheme in Europe: the Case of Photovoltaics［J］. Energy Policy, 2016, 95: 94-102.
［5］刘道华, 张飞, 张言言. 一种改进的RBF神经网络对县级政府编制预测［J］. 信阳师范学院学报, 2016, 29(2): 265-269.
LIU Daohua, ZHANG Fei, ZHANG Yanyan. A Prediction for the Preparation of County Government Based on Improved RBF Neural Networks［J］. Journal of Xinyang Normal University, 2016, 29(2): 265-269.
［6］刘焕云, 王军宁, 何迪, 等. 仿生模式识别目标自适应跟踪算法［J］. 西安电子科技大学学报, 2013, 40(4): 142-148.
LIU Huanyun, WANG Junning, HE Di, et al. Research on Biomimetic Pattern Recognition for Object Adaptive Tracking Algorithm［J］. Journal of Xidian University, 2013, 40(4): 142-148.
［7］张文鸽, 管新建, 徐清山. 水环境质量评价的模糊贴近度方法［J］. 水资源保护, 2006, 22(2): 19-22.
ZHANG Wenge, GUAN Xinjian, XU Qingshan. Water Environment Quality Evaluation Based on Fuzzy Nearness Method［J］. Water Resources Protection, 2006, 22(2): 19-22.
［8］ ZHANG M M, ZHOU D Q, ZHOU P, et al. Optimal Feed-in Tariff for Solar Photovoltaic Power Generation in China: a Real Options Analysis［J］. Energy Policy, 2016, 95(6): 181-192.
［9］ DEVECI O, KASNAKOGLU C. Performance Improvement of a Photovoltaic System Using a Controller Redesign Based on Numerical Modeling［J］. International Journal of Hydrogen Energy, 2016, 41(29): 12634-12649.
［10］ SUN C, SUN F C, MOURA S J. Nonlinear Predictive Energy Management of Residential Buildings with Photovoltaics & Batteries［J］. Journal of Power Sources, 2016, 325: 723-731.
［11］ WEERASINGHE H C, VAK D, ROBOTHAM B, et al. New Barrier Encapsulation and Lifetime Assessment of Printed Organic Photovoltaic Modules［J］. Solar Energy Materials and Solar Cells, 2016, 155: 108-116.

Method of tracing the photovoltaic maximum power point based on NN with asymmetric basis

PDF (PC)

Like

Knowledge

Abstract

Cite this article

share this article

References

Related Articles 1

Metrics

Comments

Recommended 10