Dynamic line rating forecasting

Abstract Thermal line constraints represent a cost source for power system actors, by reducing network reliability, limiting the potential renewable energy source penetration and increasing the electricity prices. They are usually faced with network reinforcements but due to their financial and administrative complexity, new solutions such as Dynamic Line Rating are being investigated. This solution is based on the fact that the real current carrying capacity of overhead lines depends on variable weather parameters. Furthermore its real time value is very often far higher than then static conservative value commonly used. The dependence of Dynamic Line Ratings on weather implies that this parameter can be forecasted, as it is done for weather dependent renewable energy production. These forecasts can be used to improve generators' dispatch, reducing overall electricity generation cost and increasing system's reliability. Three main difference are to be noticed between the renewable energy sources production forecasts and the DLR forecasts: 1) the DLR forecasts depend on several weather parameters, implying the use of a higher number of features than with PV or wind power forecast; 2) the forecasts are to be made for several points, and not a single one, due to the fact that a critical section is to be selected along the line; 3) for safety reasons, the forecasts are to be inferior to the observations most of the time, and so probabilistic forecasts are to be developed in order to provide low quantile forecasts.