Dynamic Voltage Restorer Using Intelligent CPFNN to Mitigate Abnormal Grid Voltage for Loads

An intelligent dynamic voltage restorer (DVR) is proposed in this study to rapidly balance and stabilize the load voltage during three-phase grid voltage swell, sag and imbalance conditions. Owing to contingency, unexpected load change and power system fault, the abnormal grid voltage conditions can lead to severe equipment damage and the tripping of sensitive loads. Thus, a DVR is developed to inject the controlled amount of compensation voltage to stabilize the load voltage during the abnormal grid voltage conditions. Furthermore, the dual second-order generalized integrator phase locked loop (SOGI-PLL) is utilized for the developed DVR to accurately synchronize with the power grid. Nevertheless, since the traditional proportional-integral-resonant (PIR) controllers couldn’t effectively deal with the external disturbances and parameter variations resulted in the sluggish voltage compensation of the DVR, two novel Chebyshev probabilistic fuzzy neural network (CPFNN) controllers are firstly proposed for DVR to substitute for the traditional PIR controllers to effectively improve the performance of the voltage compensation. The network architecture of the proposed CPFNN controller is derived in detailed. Finally, the effectiveness of the intelligent DVR using the proposed CPFNN controllers for load voltage improvement during three-phase grid voltage swell, sag and imbalance conditions are certified by some experimental results.