Abstract— Doubly Fed Induction Generator (DFIG) is the most widely employed generator in the Wind Energy Conversion Systems (WECS) for the production of electricity. However, despite all of its various advantages, it is extremely vulnerable to grid faults such as voltage dip since its stator is directly coupled to the grid. A voltage dip problem is one the main issues among the power quality concerns. This fault causes the flow of excessive current across both the stator and the rotor terminals, which may lead to serious damage to the generator, power converters, and DC Link capacitor. On the other hand, the current Grid Codes (GC) requires the system to stay connected to the grid during this fault condition and support it in healing its nominal voltage. This capacity of the system is known as the Low Voltage Ride Through (LVRT) capacity. For the system to achieve such capacity, appropriate protection mechanisms or controlling strategies must be utilized. Therefore, in this paper, the crowbar protection technique, PI controller, and the Adaptive Neuro-Fuzzy Inference System (ANFIS) controller are employed. Furthermore, the performance of the system employing PI, crowbar, and ANFIS is analyzed and compared under grid fault conditions, i.e., a voltage dip with a magnitude of 0.1 pu (worst case) using MATLAB/Simulink software and based on actual data obtained from Adama II wind farm. The obtained results unveil that the settling time of ANFIS for controlling the rotor currents in d and q axes (idr and iqr) and DC link voltage is 3.6 s, 3.57 s, and 3.4 s, respectively. On the other hand, the settling times of the PI controller for controlling the rotor currents in d and q axes and the DC link voltage are found to be 4 s, 3.91 s, and 45.2 s, respectively, while the crowbar protection technique’s settling times are found to be 4 s, 6 s, and 4.9 s, respectively. It is evident from the aforesaid results that the ANFIS controller provides the best performance of the three strategies since it allows both the rotor currents and the DC link voltage to return to their steady state values faster than the other two techniques, employed in this investigation.
Keywords: Voltage dip; Adaptive neuro-fuzzy inference system; Crowbar; Doubly fed induction generator; Low voltage ride through.
DOI: http://dx.doi.org/10.5455/jjee.204-1669034454