Regression‐based forecast model of induced geoelectric field

Lotz, Stefan; Heyns, M. J.; Cilliers, Pierre J.

doi:10.1002/2016sw001518

Cited by 8 publications

(10 citation statements)

References 33 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…While the predictions could be improved, the consistency of the values of Figure 1 indicate that the model is not overfitting. The correlation coefficient values are a bit lower to those obtained by Lotz et al (2017) of 0.71 and 0.69 for models predicting separate components of the horizontal magnetic field at a mid-latitude station. Wintoft et al (2015) obtain much higher correlation coefficients, although they are only considering the maximum value within each 30min window.…”

Section: Feed-forward Artificial Neural Networkcontrasting

confidence: 75%

Comparison of Deep Learning Techniques to Model Connections Between Solar Wind and Ground Magnetic Perturbations

Keesee

Pinto

Coughlan

et al. 2020

Front. Astron. Space Sci.

View full text Add to dashboard Cite

Geomagnetically induced currents (GIC) can drive power outages and damage power grid components while also affecting pipelines and train systems. 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. 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. 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. 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. The analysis techniques include a feed-forward artificial neural network (ANN) and a long-short term memory (LSTM) neural network. Here we present a comparison of both models' performance when predicting the B H component of the Ottawa (OTT) ground magnetometer for the year 2011 and 2015 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.

show abstract

Section: Feed-forward Artificial Neural Networkcontrasting

confidence: 75%

Comparison of Deep Learning Techniques to Model Connections Between Solar Wind and Ground Magnetic Perturbations

Keesee

Pinto

Coughlan

et al. 2020

Front. Astron. Space Sci.

View full text Add to dashboard Cite

show abstract

“…(2018) and Lotz et al. (2017), who also predicted ground geoelectric fields from solar wind data, is that we have approached the problem with a new tool (a recurrent neural network) and have been able to forecast GICs directly along with the geoelectric field, with the results compared to measured GICs. We have had some success, particularly with forecasting the geoelectric field, and have tried forecasting substation‐specific GICs for the first time, but there are still many problems to be addressed to turn this method into a useful forecast.…”

Section: Discussionmentioning

confidence: 99%

Forecasting GICs and Geoelectric Fields From Solar Wind Data Using LSTMs: Application in Austria

et al. 2022

View full text Add to dashboard Cite

The forecasting of local GIC effects has largely relied on the forecasting of dB/dt as a proxy and, to date, little attention has been paid to directly forecasting the geoelectric field or GICs themselves. We approach this problem with machine learning tools, specifically recurrent neural networks or LSTMs by taking solar wind observations as input and training the models to predict two different kinds of output: first, the geoelectric field components Ex and Ey; and second, the GICs in specific substations in Austria. The training is carried out on the geoelectric field and GICs modeled from 26 years of one‐minute geomagnetic field measurements, and results are compared to GIC measurements from recent years. The GICs are generally predicted better by an LSTM trained on values from a specific substation, but only a fraction of the largest GICs are correctly predicted. This model has a correlation with measurements of around 0.6, and a root‐mean‐square error of 0.7 A. The probability of detecting mild activity in GICs is around 50%, and 15% for larger GICs.

show abstract

“…Research on GIC may be divided into three areas of interest: (1) space physics and the geomagnetic response to solar activity (e.g., Lotz et al, ), (2) transformer and system response to GIC (e.g., Bolduc, ), and (3) cost of interruptions and long‐term planning (e.g., Eastwood et al, ; Oughton et al, ). Practical mitigation and long‐term planning efforts need to include contributions from all three fields.…”

Section: Introductionmentioning

confidence: 99%

Extreme Value Analysis of Induced Geoelectric Field in South Africa

Lotz

Danskin

2017

Space Weather

Self Cite

View full text Add to dashboard Cite

Extreme geomagnetic disturbances occur rarely but can have great impact on technological systems such as power supply networks. Long‐term planning for extreme events requires the estimation of event impact for occurrence periods greater than the length of observed data. With this in mind an analysis of extreme geomagnetic events observed in South Africa (middle geomagnetic latitude) is performed over four solar cycles (1974–2015). An algorithm to identify active periods with minimum SYM‐H ≤−100 nT is demonstrated. The sum of induced electric field over the course of each event is used to characterize the severity of each active period. It is found that the severity index (accumulated electric field magnitude ΣE) shares a highly linear relationship with accumulated SYM‐H over each event. The index ΣE is lognormal distributed, with tail deviating greater than lognormal, confirming heavy‐tailed occurrence. A general Pareto distribution is fitted to the tail of the distribution and extrapolated to calculate the return levels of extreme events. Return levels of once in 100 and once in 200 year events are estimated to be 9.4 × 104 mV/km min and 1.09 × 105 mV/km min, respectively. The top three events, in ascending order of severity, are the March 1989 storm, the events of late October 2003, and the April 1994 event—a long interval of coronal‐hole driven disturbances, bookended by two intense geomagnetic storms.

show abstract

Regression‐based forecast model of induced geoelectric field

Cited by 8 publications

References 33 publications

Comparison of Deep Learning Techniques to Model Connections Between Solar Wind and Ground Magnetic Perturbations

Comparison of Deep Learning Techniques to Model Connections Between Solar Wind and Ground Magnetic Perturbations

Forecasting GICs and Geoelectric Fields From Solar Wind Data Using LSTMs: Application in Austria

Extreme Value Analysis of Induced Geoelectric Field in South Africa

Contact Info

Product

Resources

About