Method Of Artificial Electromagnetic Disturbances Identification And Data Reconstruction

Geomagnetism is a discipline based on observation. Geomagnetic observation data possess academic significance and indispensable strategic resource. High-precision geomagnetic data provides condition for researches of seismology, geology,exploration and geomagnetism.Since 2008,Chinese digital geomagnetic precursory observation network has been built, and large quantities of High-precision geomagnetic data were accumulated. However, as China is experiencing the period of rapid social and economic development, artificial electromagnetic noise causes serious disturbances to geomagnetic observation data. Therefore, it is an urgent task to identify various type of artificial electromagnetic disturbances for the quality improving of geomagnetic observation data.In this paper, we analyze the time and frequency characteristics of geomagnetic field and then proposed methods to identify artificial electromagnetic disturbances and rebuild the disturbed data.The main contributions are summarized as follows:1. FFT transform on one-second data. The sampling of one-second data is highest in all data; therefore we use it to conduct researchs. Based on FFT-transform of one-second data, the amplitude of frequency spectrum was obtained.2. Research on spatial and temporal distribution of spectrum. Geomagnetic field changes over time and frequency regularly. It depends on local-time and an obvious day-night asymmetry exists. The characteristics of geomagnetic spectrum show 27 solar-cycle recurrence and significant seasonal variation in temporal distribution and regular changes with latitude and longitude in spatial distribution. Meanwhile geomagnetic field increases with the decrease of frequency. We propose a method to identified artificial electromagnetic disturbance based on the characteristics both in time and frequency domain.3. Research on parameters determination. We analyze the change of parameters in different signal-to-noise ratio (SNR) standard and then get the final parameters considered actual observation situation. Eventually,the electromagnetic disturbances were recognized different from normal geomagnetic variations. We divide disturbances into three types,environmental disturbances,instrument troubles and shocking disturbances.4. Research on methods of data reconstruction. We use space weighting method and Back-Propagation Network to rebuilt missing or disturbed data and obtained good result.