A Novel Hybrid Model for Dynamic Line Rating Using Artificial Neural Networks
A. Alarcon, I. Sanz, A. Lostale, P. Inda, M.G. Cañete, S. Borroy, A. P. Talayero CIRCE Centro Tecnologico, Zaragoza, Spain 2026-06-27

Abstract

Dynamic Line Rating (DLR) is a smart gridtechnology that calculates the real-time, maximum power capacity (ampacity) of transmission lines based on current environmental conditions. DLR allows higher, safe power flows, making a better use of the line’s actual capacity and thus allowing for the integration of more renewables.

This document presents a hybrid architecture that combines the physical IEEE 738 standard with advanced deep learning techniques, Long Short-Term Memory Neural Networks (LSTM) and Physics‑Informed Neural Networks (PINNs), to predict the thermal ampacity of overhead transmission lines.

The architecture of the neural networks is capable of learning complex patterns from historical climatological data. This enables the model to capture the temporal and stochastic evolution of environmental conditions. Through this data‑driven learning process, the DLR system can significantly reduce the reliance on field sensors, and in some cases even operate with minimal or no sensing infrastructure, by inferring the necessary conditions directly from historical and meteorological inputs.

While the traditional IEEE 738 approach is deterministic and static, providing only point estimates under fixed conditions, modern power system operation demands dynamic forecasts, uncertainty management, and physical consistency, motivating the integration of physical models with deep learning.

Key words: Ampacity, Climatological Conditions, Sensorless, Grid congestions.

References

[1] IEEE Standard 738-2012, ¨IEEE Standard for Calculating the Current-Temperature Relationship of Bare Overhead Conductors,¨IEEE Power and Energy Society, 2013.

[2] M. Raissi, P. Perdikaris, and G.E. Karniadakis, ”Physics-informed neural networks: A deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations,”Journal of Computational Physics, vol. 378, pp. 686-707, 2019.

[3] S. Hochreiter and J. Schmidhuber, ”Long short-term memory,”Neural Computation, vol. 9, no. 8, pp. 1735-1780, 1997.

[4] A. Michiorri et al., ”Forecasting for dynamic thermal rating,”IEEE Transactions on Power Systems, vol. 30, no. 2, pp. 663-673, 2015.

[5] R. Sims et al., ¨Integration of renewable energy into present and future energy systems,” IPCC Special Report on Renewable Energy Sources and Climate Change Mitigation, Cambridge University Press, 2011.

[6] V. Terzija et al., ”Wide-area monitoring, protection, and control of future electric power networks,”Proceedings of the IEEE, vol. 99, no. 1, pp. 80-93, 2011.