Fault detection, classification in multiterminal HVDC transmission system with MC-SVM
What is it about?
This paper proposes a protection scheme for the DC lines in VSC-HVDC system using transient-voltage. The proposed protection scheme, which consists of a protection methodology is based on the MC-SVM. It is a machine learning based approach. For the training purpose of MC-SVM four types of faults are created in two different sections at every 10 km of the transmission line and the data is measured. By employing the proposed protection methodology, fault detection and classification with high selectivity and robustness against fault resistance and disturbance are achieved.
Why is it important?
Faults are bound to occur in the system therefore fault identification and classification play a vital role in achieving reliable operation. Moreover this becomes even more complicated for multi-terminal HVDC transmission system because, it mainly connects large offshore wind farms to the load centre at different generating frequencies. When the fault occurred, the generation will be wasted until the system is restored. In order to facilitate fault identification and classification an algorithm is required which can act fast without sacrificing precision and reliability. Till now lots of work has been carried out for the protection of HVAC transmission system, however HVDC transmission system needs more attention towards fault identification and classification. Therefore, for HVDC transmission system a new protection methodology with high selectivity and reliability is essential.