We introduce the empirical mode decomposition algorithm and applied low-frequency filtering and high-frequency noise reduction to the waveform of the electricfield changes recorded in 1-ms segments. This algorithm gr...
详细信息
We introduce the empirical mode decomposition algorithm and applied low-frequency filtering and high-frequency noise reduction to the waveform of the electricfield changes recorded in 1-ms segments. This algorithm greatly improved the accuracy of the peak time extraction and the number of pulses in the low and very lowfrequency band, which enhanced the accuracy in positioning the pulse of the electricfield change. Compared with the previous algorithm, the algorithm can significantly reduce the time error, giving a better positioning result. With a time error estimate of 100 ns and a limit of goodness of fit less than 5, the number of pulse locations is increased by nearly 7 times. The goodness-of-fit distribution of the pulse location results had a normal distribution, and the 95% confidence interval of goodness of fit was 0-4;the corresponding positioning space error was <60 m. The continuity of the lightning channel was significantly improved, and the development characteristics and fine structure of the lightning channel were clearly distinguished. The low-frequencyelectricfielddetectionarray system gave detailed positioning results for a bolt from the blue lightning strike. By comparing the results from the low-frequencyelectricfielddetectionarray with the actual lightning strike point, we objectively demonstrated the positioning performance of the new algorithm. The system gave positioning results for the lightning for all seven return strokes. The maximum horizontal distance between the locating point and the real lightning strike point was 57 m, the minimum horizontal distance between them was 3 m, and the mean distance was 27 m.
暂无评论