An Interval Prediction Method for Day-Ahead Electricity Price in Wholesale Market Considering Weather Factors
Xin Lu, Jing Qiu, Gang Lei, Jianguo Zhu
Abstract
Accurate prediction of electricity prices under uncertainties is an important and challenging problem for all electricity market participants. This article proposes a novel generative model-based prediction interval construction method for day-ahead electricity prices. A conditional time series generative adversarial network is proposed to generate realistic and diverse electricity price scenarios. With these generated price scenarios, prediction intervals can be combined. After that, a threshold select machine is proposed to truncate the threshold of random noise input to adjust the quality of prediction intervals, balancing the reliability and sharpness. Finally, shortwave irradiance, wind speed, and temperature are taken into account in the threshold select machine, further improving the reliability and the sharpness of the prediction intervals. Case studies verify the effectiveness and superiority of the proposed method.