Model-based diagnosis for sequential shunt faults in HVDC transmission lines

This work proposes a model-based diagnosis for detecting and locating sequential conductance deteriorations faults on-line in monopolar high-voltage direct current (HVDC) transmission lines, if only measurements are available at the ends of the line. The basis of the diagnosis algorithm is a recursive procedure of a distributed adaptive observer scheme named the shunt fault estimator scheme (SFES), which monitors the line by identifying and locating any single shunt. By considering the equivalence between the transmission line model with one and two faults, two outcomes are shown: (1) the lack of isolability in steady-state for two simultaneous shunt faults and (2) the ability of the SFES to solve the two faults' scenario recursively if the shunt faults occur sequentially. Thus, by using the equivalence relationships between the parameters for one and two faults, a location algorithm can be designed through the SFES that provides accurate and fast parameter estimations for the two faults that can be applied on-line. Its performance is studied with numerical simulations and with synthetic data obtained from a deteriorated line 300 x 103 [m] long. The location time in this case was fewer than approximately 4 [ms] that displays the algorithm potentiality in a practical diagnosis scenario.