摘要
针对现有太阳能电池板转动方式相关研究中,忽略转动能量损耗以及不能适应特殊天气变化的缺陷,提出了一种基于转动特征选择和电力视觉技术的太阳能电池板转动策略.该策略分析了能够影响转动的相关特征并进行特征选择,在此基础上采用RBF网络训练相关参数确定特征权重,且提出了基于深度卷积神经网络的光照时长预测方法.实验结果表明,所提出的转动策略能够从整体上提高太阳能电池板的转动收益,并且能更好地适用于各类特殊天气的变化,通过与同类方法对比表明所提出的转动策略可以提升整体转动效率20%以上.
Aiming at the problem that the loss of rotation energy and the special weather changes were ignored in the relevant research on rotation mode of solar panel,a solar panel rotation strategy based on rotation feature selection and power vision technology was proposed.In this strategy,relevant features which may affect rotation was considered and feature selection was implemented,the RBF network was used to train relevant parameters for the determination of feature weight,and a method for sunlight duration prediction based on deep convolution neural network was proposed.The results of experiments show that the as-proposed rotation strategy can improve the overall rotation income of solar panels,and it can also be better adapted to various special weather changes and can improve the overall rotation efficiency by more than 20%in comparison with similar methods.
作者
于舜
侯荣旭
郭朋伟
张聿博
夏炎
吴欣怡
YU Shun;HOU Rong-xu;GUO Peng-wei;ZHANG Yu-bo;XIA Yan;WU Xin-yi(School of Information,Shenyang Institute of Engineering,Shenyang 110136,China;College of Information Science and Technology,Northeast Normal University,Changchun 130024,China)
出处
《沈阳工业大学学报》
CAS
北大核心
2023年第3期336-341,共6页
Journal of Shenyang University of Technology
基金
国家自然科学基金联合基金项目(U1811261)。
