Prediction interval estimation of dynamic thermal rating considering weather uncertainty

Uncertainty and randomness of weather bring challenges to the accurate estimation of dynamic thermal rating (DTR). To reduce the impact of weather factors on the DTR predictions and to further enhance transmission line ampacity. An ensemble empirical mode decomposition quantile regression temporal convolutional network (EQRTCN)-based interval prediction model is proposed for DTR. Different from the existing DTR techniques, the model overcomes the effects of weather non-smoothness and uncertainty on prediction intervals with low quality by ensemble empirical mode decomposition (EEMD) and quantile regression theory. In this work, a k-means clustering algorithm is first used for the time boundary condition division by considering the coupling between line currents and weather parameters. Then, based on different time boundary conditions, the DTR interval prediction model is proposed by using EEMD, quantile regression theory and temporal convolutional networks. Finally, the proposed improved method for ampacity calculation is incorporated in the EQRTCN to calculate the prediction interval of DTR. The model simulation results demonstrated that the proposed DTR prediction framework effectively reduced the uncertainty introduced by meteorological parameters in interval predictions and provided safe guidance for power dispatch.