Application and evaluation of time-domain differential protection based on bayesian inference to HVDC lines

This paper presents the application of a time-domain differential protection algorithm to protect high-voltage direct current (HVDC) lines. The method discussed in this article was validated and proved fast and effective in protecting alternating current (AC) transmission lines. It applies Bayesian Inference (BI) to the current sampled values to determine internal fault occurrence probabilities over time. The authors performed fault simulations on a bipolar ± 500 [kV] version of Cigré's benchmark model for HVDC control studies in the Alternative Transients Program (ATP) and generated multiple cases, varying fault types, application points, resistances, and start instants to evaluate the proposed method's performance. Besides, the authors also tested the algorithm against situations featuring noisy measurements and terminal synchronisation failures. The simulation results were re-sampled to sampling rates typically found in commercial device and used for implementing the protection in Matlab. The paper presents the test results for the discussed method, evaluating its fault detection speed and reliability.