Recently, the issue of protection in active distribution networks has been greater than before due to the mushrooming use of photovoltaic (PV) generation units. PV systems coupled to power electronics devices and their
control strategies causes significant changes to fault current characteristics, leading to maloperation of fault
identification schemes. Therefore, this paper proposes a robust, self-adaptive fault identification scheme that
incorporates the presence of PV systems and considers the effect of uncertainty arising due to their high penetration levels. The real-time voltage and current measurements are captured locally by advanced one-end
measuring facilities, considering the uncertain temperature and irradiation conditions, of solar generation
units. First, the proposed scheme employs wavelet transform, statistical alienation concepts, and fuzzy logic to
identify the fault criteria. Then, it uses fuzzy concept-based system features to obtain optimal protection
scheme setting. For simulation analysis, a 10 MW solar farm-connected to a real-world active distribution system
is emulated in PSCAD-EMTDC simulator to create intensive study cases, and the proposed scheme is implemented
in MATLAB software. Several fault parameters are examined in the scheme, comprising different fault types, fault
locations, and transition resistances. Also, the uncertainty arising from PV plants and different loading conditions
are taken full consideration in the developed scheme. Simulation results have verified the validity of fault
identification scheme; nonetheless, the computational burden is kept reasonable. The proposed scheme is
investigated under white gaussian and impulsive noise. Furthermore, sensitivity analysis and comparative study
are conducted in terms of accuracy, high decision speed, and sampling frequency variation. The proposed scheme
succeeds to detect and classify internal faults for both on- and off-grid operation modes and differentiate them from
external ones in the case of on-grid mode. The proposed scheme has dominance performance, including high
speed fault detection within a half cycle and provides accuracy greater than 98%. |
