Abstract
In traditional pantograph-catenary arc mathematical models, the relationship function between high-speed airflow and the voltage gradient contains certain inaccuracies. Furthermore, fault identification based on voltage or current often involves complex feature extraction and algorithmic processes. To address these issues, this paper introduces a new mathematical model and presents the concept of port impedance. First, based on the Habedank model, this paper combines the offline distance and airflow velocity fitting function with the existing relationship function between voltage gradient and arc current, deriving a mathematical arc model with airflow velocity and arc current as the primary influencing factors. Next, the correctness of this arc model is validated through simulations and experiments. The arc model simulation is performed using PSCAD software, and a corresponding experimental platform is built in the laboratory. Based on this, the Spearman correlation coefficient between the simulation and experimental data is calculated, showing a high correlation between the simulation data and the experimental data. Finally, the concept of port impedance is introduced, and the feasibility of arc fault identification based on port impedance is analyzed using the validated arc model. The research results indicate that port impedance can transform fault diagnosis problems into linearly separable problems, thereby avoiding the high-dimensional feature extraction and complex algorithms involved in arc fault diagnosis using current or voltage. Even without optimization, the original support vector machine classification model achieves an identification accuracy of up to 99%.