Risk-Averse Stochastic Dynamic Line Rating Models

Abstract

Static line rating (SLR), which is conventionally used in operation, not only results in a conservative usage of the capacity of overhead lines, but also fails to accurately address the overload risk. In this paper, using quantile regression (QR) and superquantile regression (SQR) methods, two models are proposed to predict dynamic line rating (DLR) of overhead conductors in operational applications with very short-term horizons. The proposed methods model statistical properties of time evolution of conductors considering the conductor thermal inertia to cope with situations with higher time resolutions for enhanced capacity usage. To address the overload risk due to forecast uncertainties of weather-related parameters, the proposed models are reformulated as risk-based constraints and utilized as QR and SQR-based DLR. The developed constraints are fully parametric and readily applicable to optimization problems and are verified through an optimal power flow (OPF). Results of examining the proposed models on the RTS test system confirm their efficiency in terms of better utilization of conductor capacity, increased energy transfer, and reduced risk levels.