Dynamic Line Rating Using Numerical Weather Predictions and Machine Learning: A Case Study

Abstract

In this paper a dynamic line rating experiment is presented in which four machine learning algorithms (Generalized Linear Models, Multivariate Adaptive Regression Splines, Random Forests and Quantile Random Forests) are used in conjunction with numerical weather predictions to model and predict the ampacity up to 27 hours ahead in two conductor lines located in Northern Ireland. The results are evaluated against reference models and show a significant improvement in performance for point and probabilistic forecasts. The usefulness of probabilistic forecasts in this field is shown through the computation of a safety-margin forecast which can be used to avoid risk situations. With respect to the state of the art, the main contributions of this paper are: an in depth look at explanatory variables and their relation to ampacity, the use of machine learning with numerical weather predictions to model ampacity, the development of a probabilistic forecast from standard point forecasts and a favo urable comparison to standard reference models. These results are directly applicable to protect and monitor transmission and distribution infrastructures, especially if renewable energy sources and/or distributed power generation systems are present.