?
3‐D Finite Difference Time Domain Simulation of Lightning Strikes to the 634‐m Tokyo Skytree
Waveforms of vertical electric field observed at distances of 27, 57, and 101km from the 634-m tall Tokyo Skytree (Japan) struck by lightning have been reproduced using the 3-D finite difference time domain method. Optically observed 3-D lightning channels and currents directly measured in the tower were used. Distribution of current along the lightning channel was represented using the transmission line model, assuming a constant propagation speed v=0.5c. Simulations with commonly assumed vertical channel did not allow us to reproduce the observed early (within 5 s or so) field zero crossing when the 1-D speed was set to be constant and equal to the 3-D speed (v=0.5c) used for the actual-channel-geometry case. It was, however, possible to achieve a good match for a vertical channel when the vertical component of 3-D speed was evaluated for each nonvertical channel segment and used in the corresponding range of heights.
Plain Language Summary Since tall towers are often struck by lightning, direct lightning current measurements are usually performed on such towers. Also, lightning strikes to tall towers have been used to test lightning locating systems in terms of their detection efficiency, location accuracy, and current estimation errors. Hence, it is important to study the interaction of lightning with towers and associated electromagnetic fields. In this letter, waveforms of vertical electric field observed at distances of 27, 57, and 101km from the 634-m tall Tokyo Skytree struck by lightning have been reproduced using the 3-D finite difference time domain method. Optically observed 3-D lightning channels and currents directly measured in the tower were used. Distribution of current along the lightning channel was represented using the transmission line model, assuming a constant propagation speed v=0.5c. Simulations with commonly assumed vertical channel did not allow us to reproduce the observed field zero crossing when the 1-D speed was set to be constant and equal to the 3-D speed (v=0.5c) used for the actual-channel-geometry case. It was, however, possible to achieve a good match for a vertical channel when the vertical component of 3-D speed was evaluated for each nonvertical channel segment and used in the corresponding range of heights.