Abstract—Damage of electric power facilities caused by typhoons is one of the most common meteorological disasters in Japan. The electric power supply is sometimes cut off in wide areas for a long time by typhoons, which brings an undesirable effect on society. To ensure the speedy restoration of the electric power supply, it is important to predict the amount of damage accurately for an approaching typhoon. This paper presents an improvement of Gaussian process (GP)-based predictor of electric power damage caused by typhoons. The electric power facilities have been maintained so that they are robust to the disasters, and the robustness changes with time. Therefore, the proposed predictor uses the information of the arrival date of typhoons as well as the typhoon weather information. Simulation results based on actual data show that the proposed predictor improves the accuracy of prediction compared with the conventional GP-based predictor.

Index Terms—prediction, damage by typhoon, electric power systems, gaussian process, time-varying characteristics

Cite: T. Hachino, S. Okubo, H. Takata, S. Fukushima, and Y. Igarashi, "Improvement of Gaussian Process Predictor of Electric Power Damage Caused by Typhoons Considering Time-Varying Characteristics," International Journal of Electronics and Electrical Engineering, Vol. 3, No. 4, pp. 263-268, August 2015. doi: 10.12720/ijeee.3.4.263-268