论文部分内容阅读
研究我国南海地区（北纬0°到25°，东经105°到120°）电离层F层临界频率月中值变化特点及利用人工神经网络预报月中值的技术．文章首先给出这个海区上空F2层临界频率月中值随太阳活动水平、地理纬度、季节和昼夜变化规律．结果指出电离层临界频率随纬度分布除出现驼峰现象外，在4和10月份还存在大范围的电离层气泡区域。在这个区域电离层密度出现中间稀周围高特点；电离层临界频率的分形分析得到其相关维为2．7。其次，我们选用5个因子作为网络输入，用电离层观测资料训练网络，建立了我国南海地区电离层模式。利用这个模式预测的电离层临界频率与实际测量的符合较好，可以满足应用要求。
The characteristics of the mid-monthly variation of the critical frequency of the F-stratosphere in the South China Sea (latitude 0 ° to 25 °, longitude 105 ° to 120 ° E) and the technique of using Artificial Neural Network to predict the median of the moon are studied. The paper first gives the monthly variation of the critical frequency of F2 layer with the solar activity level, geographical latitude, season and diurnal variation. The results indicate that in addition to the hump phenomenon, the critical ionospheric frequency distribution along with the latitudinal distribution also shows a large range of ionosphere bubble regions in April and October. In this region, the ionospheric density is characterized by middle dilute high and the correlation dimension is 2.7 by the fractal analysis of the ionospheric critical frequency. Secondly, we choose 5 factors as the network input, train the network with ionospheric observations, and establish the ionospheric model in the South China Sea. The predicted ionospheric critical frequency using this model is in good agreement with the actual measurement to meet the application requirements.