三亚地区电离层foF2的混沌特性分析及其预测研究
Analysis of chaotic features and prediction of ionospheric foF2 in Sanya
投稿时间:2019-02-17  修订日期:2019-02-17
DOI:
中文关键词: 电离层foF2  混沌  神经网络  预测
英文关键词: ionosphere foF2  chaos  neural network  prediction
基金项目:国家自然科学基金项目(41474135,41474134)
作者单位E-mail
朱正平 中南民族大学 zpzhu2007@sina.com 
阮鹏飞 中南民族大学 febfeiniao@163.com 
摘要点击次数: 67
全文下载次数: 0
中文摘要:
      摘 要 本文利用三亚台站2013年电离层foF2观测数据讨论了电离层foF2的混沌特性及其预测。采用改进的C-C算法确定时间延迟和嵌入维数。然后计算最大李雅普诺夫指数,从定量的角度印证foF2时间序列具有混沌特性。最后,基于RBF神经网络的方法对foF2参量进行短期预报,将预报结果与Volterra模型、IRI模型和实测数据进行对比。结果表明,采用RBF神经网络法进行预测是成功的,相比于国际参考电离层模型有较大提高,较Volterra模型也有一定提升。在一定预测尺度范围内,RBF神经网络预测结果较为准确,预测误差较小,超出该预测范围,预测效果将变差。
英文摘要:
      Abstract This paper discusses chaotic features of critical frequency of the ionospheric F2 layer (foF2) and its prediction based on the observation data of foF2 in Sanya, 2013. The improved C-C algorithm is adopted to determine the time delay and embedding dimension. Then the maximum lyapunov index is calculated to verify the chaotic property of foF2 time series quantitatively. The method based on RBF neural network is used for short-term forecast of foF2 parameters and the forecast results were compared with the Volterra model, the IRI model and the measured ones. It shows that the prediction using RBF neural network is successful. The method of using RBF neural network is superior to IRI model and Volterra model. Predicted results are more accurate and less prediction errors within predictable range using RBF neural network. Beyond the range, the effect will be worse.
View Fulltext   查看/发表评论  下载PDF阅读器
关闭