Abstract

This paper presents a new simulation approach for electromagnetic transient analysis on lossy transmission lines, based on combining the Radial Basis Function-based Finite-Difference Time-Domain (RBF‑FDTD). The proposed method overcomes limitations of the traditional FDTD scheme, which relies on Taylor-series-based finite-difference operators, by employing RBFs to construct more flexible and accurate numerical approximations. A new algorithm is introduced to determine the optimal RBF shape parameter, enabling a balanced trade-off between accuracy, numerical stability, and computational efficiency. The algorithm minimizes local approximation error through the use of a weight matrix and Taylor expansion, allowing the optimal parameter to be identified without requiring an exact reference solution. The RBF‑FDTD method is validated through multiple transmission-line models under different loading conditions, including resistive, parallel RC, RL, and parallel RLC terminations. Simulation results demonstrate that the method accurately reproduces transient waveforms, closely matching reference solutions in terms of amplitude, reflection timing, and resonant characteristics, while maintaining numerical stability in the presence of losses. Compared with the conventional FDTD method, RBF‑FDTD shows clear advantages in modeling nonuniform or distributed-parameter structures where high accuracy in spatial derivative approximation is required. These findings confirm the potential of RBF‑FDTD as a powerful and efficient tool for transient analysis in power systems, and highlight its applicability to more complex configurations such as branched transmission networks, grounding systems, and lightning-induced transients.