An intelligent algorithm for fault location on VSC-HVDC system

- A novel intelligent algorithm for fault location on VSC-HVDC system is proposed.
- This method with makes full use of frequency, time and energy. And it capture the fault features via Hilbert-Huang Transform.
- The bat algorithm optimizes the parameters of the model. High-frequency variance contribution rate is adopted to identify the fault area.
- The expected accuracy of the proposed method is ±500 m.
- The model is also applicable for hybrid transmission systems.

This paper proposes a novel intelligent algorithm for fault location on the voltage sourced converter-high voltage direct current (VSC-HVDC) system. This method with single-ended measurement makes full use of frequency, time and energy to capture the fault features via Hilbert-Huang Transform (HHT). The time delay, characteristic frequency, energy attenuation and high-frequency energy are used as the input of ε-support vector regression (SVR) to get fault distance. Then bat algorithm (BA) optimizes the parameters of the model with cross validation comparing with other optimization algorithm. Furthermore, high-frequency variance contribution rate (HVCR) is adopted to identify the fault area. The VSC-HVDC simulation system is constructed to verify the reliability and accuracy of the method and the expected accuracy of the proposed method is ±500 m. The results demonstrated that the proposed method still has reliability and accuracy for hybrid transmission line with a small amount of data.