Thunderstorm location using electrostatic field sensor network

Thunderstorm is an atmospheric phenomena characterized by the presence of lightning. For simplicity, this charge transfer has been modelled using non-linear point charge models that allow location and identification of electrical characteristics of the phenomena. In this work, the location problem is seen like a non-linear optimization problem. In This way the main goal is to evaluate methods that improve the performance of the classical methods used by other investigations from the point of view of variables uncertainties, convergence and computing time under uncertainty conditions and measurement error. Three optimization methods are compared: two of them based on gradient information and one on heuristic principles. These methods were chosen by analysing the objective function and the constraints of the problem.