Power Quality Enhancement Using Dynamic Voltage Restorer With Echo State Network and Type-2 Fuzzy Logic Control

This research proposes an Echo-State Network (ESN) driven strategy to improve Power Quality (PQ) in utility grids through Dynamic Voltage Restorer (DVR) implementation. With the growing integration of nonlinear or critical loads, PQ degradation has become a pressing concern. The ESN-based algorithm efficiently estimates fundamental voltage quantities from a disturbed grid and supports PQ improvement. The proposed approach integrates an Echo State Network (ESN) for real-time estimation of fundamental voltage components and an adaptive Interval Type-2 Fuzzy Logic (IT2 FLC) controller for DC link voltage stabilization. This dynamic controller mitigates voltage stress and maintains a robust DC voltage level, crucial for generating the reference load voltage. The controller gains are optimized using the bio-inspired Gorilla Troops Algorithm (GTA), which enables a low computational burden and enhanced adaptability to grid disturbances. Additionally, it is noteworthy that the proposed technique reduces the values of settle and rise time (0.15s, 0.19s), overshoot and undershoot (3.28%, 3.56%), and recovery time to 0.421s. These statistics indicate that it outperforms the traditional SRF-PI and LSM-IT1 FLC strategies. The hybrid ESN and GTA IT2 FLC framework allows online gain tuning, ensuring sinusoidal load voltage and a significant reduction in total harmonic distortion from 12.19% to 2.74%. The proposed method delivers superior performance even in disturbed grid conditions, offering fast rise time, minimal overshoot, and less steady-state errors compared to conventional SRF-PI and LMS-Fuzzy Logic Controller (LMS-FLC). The findings obtained from a scaled-down prototype of DVR highlight the approach's effectiveness.