![]() For example, recurrent neural networks such as long short-term memory (LSTM) techniques are used to "remember" parameters from earlier times that have a strong influence on the output features. More complex neural network architectures can be used to improve the predictions for time-series data. (2015), they found reasonable predictions of the timing of intense fluctuations, with less accuracy as the storm evolved. They developed separate models for the north and east components of the geomagnetic field and found that fluctuations in the eastward component are more dependent on the interplanetary magnetic field (IMF) B z. Lotz and Cilliers (2015) developed a neural network based model using solar wind and IMF inputs and dB/ dt measurements at a Southern hemisphere mid-latitude station as outputs. While many studies of GICs focus on high magnetic latitudes (>60°) that lie under the auroral oval, it has been shown that mid- (50°–60°) and low- (<50°) latitude regions are also at risk ( Gaunt and Coetzee, 2007 Ngwira et al., 2008 Pulkkinen et al., 2010 Oliveira et al., 2018). Their models generally predict the timing of GICs caused by sudden impulses well, even when they train the model using only ACE magnetic field measurements. (2015) developed models using Elman neural networks to predict the 30-min maximum of dB H/ dt (horizontal component of dB/ dt) from ACE solar wind and magnetic field measurements. Machine learning based models have the potential for providing efficient, computationally inexpensive forecasts. Such models are computationally expensive and take longer time to run, posing challenges for their use as a forecasting tool. Physics based models are used to determine magnetic field fluctuations, but high resolution models are needed to obtain the spatially localized variations ( Welling et al., 2019). They theorize that it could be due to “the mapping of magnetospheric currents to local ionospheric structures,” but indicate that further study is needed. (2018) studied two storms during which intense dB/ dt peaks occurred and indicated that substorms appear to be the driver of GICs, but state that it's not clear how the widespread features of substorms lead to localized peaks in dB/ dt. Thus, measurements of dB/ dt using ground magnetometers are used as a proxy for studying GICs. The geoelectric field is driven by temporal changes in the magnetic field and the local geology. However, neither measurements of GICs nor the geoelectric field are readily available. ![]() The intensity of GICs is determined by the strength of the geoelectric field. Thus, the ability to forecast GICs is of significant interest to the space weather community, industry partners, and national interests. Geomagnetically induced currents (GICs) are one of the most significant space weather effects due to their potential to damage the power grid and can cause widespread, long-term power outages. Here we present a comparison of both models' performance when predicting the B H component of the Ottawa (OTT) ground magnetometer for the year 20 and then when attempting to reconstruct the time series of B H for two geomagnetic storms that occurred on 5 August 2011 and 17 March 2015. The analysis techniques include a feed-forward artificial neural network (ANN) and a long-short term memory (LSTM) neural network. We apply two techniques for time series analysis to study the connection of solar wind and interplanetary magnetic field properties obtained from the OMNI dataset to the ground magnetic field perturbations. We are developing a set of neural networks to predict the east and north components of the magnetic field, B E and B N, from which the horizontal component, B H, and its variation in time, dB H/ dt, are calculated. While GIC data is not readily available, variations in the magnetic field, dB/ dt, measured by ground magnetometers can be used as a proxy for GICs. Developing the ability to predict local GICs is important to protecting infrastructure and limiting the impact of geomagnetic storms on public safety and the economy. Geomagnetically induced currents (GIC) can drive power outages and damage power grid components while also affecting pipelines and train systems. 3Department of Physics and Geophysical Institute, University of Alaska Fairbanks, Fairbanks, AL, United States.2Department of Electrical and Computer Engineering, University of New Hampshire, Durham, NH, United States.1Department of Physics & Astronomy and Space Science Center, University of New Hampshire, Durham, NH, United States.Keesee 1 *, Victor Pinto 1, Michael Coughlan 1, Connor Lennox 2, Md Shaad Mahmud 2 and Hyunju K.
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |