Overhead transmission lines dynamic rating estimation for renewable energy integration using machine learning

Abstract

The increasing electrical power generation from renewable energy resources poses new challenges to the electrical grid. These challenges are mainly related to the inconsistent nature of renewable energies and the burden on transmission lines due to the increasing generated power. In this paper, the correlation between renewable energy generation and the transmission capability of electrical power lines is studied. In order to facilitate the accommodation of renewable energies the actual dynamic line rating (DLR) is estimated based on actual weather conditions. To achieve this purpose a new DLR estimation method using machine learning is presented. Line rating depends on various meteorological factors (wind speed, wind direction, ambient temperature, and solar radiation). A machine-learning model was trained using Random Forest (RF) regressor to estimate the rating of a transmission line based on these meteorological factors. The data used for model training is actual measurements of weather conditions, conductor temperature, and conductor current. Results, where the trained model is used to estimate the rating of an ACSR Lynx conductor with a thermal limit of 80°C, are presented.

Keywords

• Renewable energy
• Dynamic line rating
• Machine learning
• Random forest regressor