A Novel Neuro-Based Approach for Predicting the Location of Cloud-to-Ground Lightning Return Strokes
S. A. Emamghoreishi, R. Moini, S.H.H. Sadeghi, Mohammad Bagher Menhaj · 2001
In this paper, an attempt is made to estimate the cloud-to-ground lightning location by neural network algorithms. A two layer feed forward neural network with levenberg-marquardt learning law was employed. To produce required data for training the network, two common models TL and MTLL were utilized to compute the lightning electromagnetic fields. Simulation results demonstrate that the network can estimate the location of return stroke channel with a maximum error of 1 km in the range of 1 to 80 kms. Also, the histogram of distance error shows that in most cases the error is lower than 500 meters for both models.