کد مقاله | کد نشریه | سال انتشار | مقاله انگلیسی | نسخه تمام متن |
---|---|---|---|---|
1776553 | 1021684 | 2014 | 15 صفحه PDF | دانلود رایگان |
• First attempt to predict H component over Indian sector using NN model.
• Optimum choice of input parameters for NN model is identified.
• Model captures diurnal, seasonal, solar activity and latitudinal aspects.
• Model efficiency is high for quiet and storm periods.
Present work is the first attempt to predict horizontal component of earth's magnetic field (H) and range in H (ΔH) over Indian sector by considering the stations, namely, Trivandrum, Pondicherry, Visakhapatnam, and Nagpur, using the concept of neural network (NN). Through training procedure, solar flux (F10.7), latitude, longitude, day of the year, local time, Ap index, IMF Bz, and ion number density are identified as the optimum choice of input parameters, whereas the inclusion of solar wind pressure and velocity has not significantly improved the performance of the model. Thus an appropriate neural network model, NSSHC has been developed with 12 hidden neurons and 500 iterations to predict H component and range in H (ΔH) during the period 1996–2001, to capture diurnal, seasonal, latitudinal, magnetic and solar activity effects.
Journal: Journal of Atmospheric and Solar-Terrestrial Physics - Volume 121, Part B, December 2014, Pages 206–220