Publicações

Thunderstorms cause a significant number of faults in energy distribution systems. The possibility to forecast the thunderstorm frequency in the summer period with a reasonable anticipation (a few months) would be of great importance to support preventive actions for protection and recomposition of the electrical system, improving its performance and reliability. Thunderstorm day monthly records obtained in the last 50 years in the city of Campinas are used to develop a forecast method to predict the lightning activity with at least a month in advance in the area of operation of the Elektro power electric company in the state of São Paulo, in the Southeast region of Brazil. It was found that the thunderstorm day monthly records during the summer period are not correlated with large scale meteorological parameters that are considered to be associated to the convective activity in the South America such as the El Niño Southern Oscillation (ENSO), the South Atlantic Ocean surface temperature anomaly, the Pacific Decadal Oscillation (PDO) and the international sunspot number related to the solar activity, while they are good correlated with previous thunderstorm days if conveniently filtered to avoid high frequency changes. Based on this fact, a prediction method with a one month in advance was developed with a reasonable level of success. For larger periods the confidence of the results decreases significantly. The preliminary results of the method are supported by comparisons with lightning data obtained by the Brazilian Lightning Detection Network (BrasilDat) and interruptions data in the distribution network for the last years, indicating its high potential to support operation management.