Adaptive LMS and Optimized Fractional Order-Based Control for DVR Performance Enhancement

To protect voltage-dependent loads from disturbances, a Dynamic Voltage Restorer (DVR) has been evaluated for its performance in ensuring the seamless operation of a three-phase distribution system. The DVR’s performance has been analyzed under various voltage disturbance conditions at the source. Addressing the drawbacks of the classic fixed step-size least mean square algorithm based on geometric algebra, which is unable to efficiently lower steady-state error and attain a faster rate of convergence, an enhanced Logarithmic Function-based Variable Rate Geometric Least Mean Square (LFVRG–LMS) is suggested for the load reference voltage estimation. Secondly, balancing of adaptive DC voltage amplitude across capacitor plays a deciding role for the restoration of load voltage. The Fractional Order PID controller is utilized to maintain the desired voltage of the DC link. The gain controllers of FOPID are fine-tuned by Zebra Optimizer (ZO), and their capabilities are examined under dynamic response to substantiate the method’s superiority. The proposed work provides a performance evaluation study with different competitor optimizers like Equilibrium Optimizer (EO) and Bald Eagle Optimizer (BEO). The performance evaluators, like rising, settling time, maximum overshoot, and recovery timing, are chosen to showcase the potency of Zebra Optimizer (ZO). The results inferred that the rising time (0.124 s), quick settling time (0.24 s), high overshoot (11%), undershoot (8.76%), and recovery time (0.58 s) are obtained with the ZO-based FOPID controller compared to the other optimization methods. Thus, the reduced overshoot, lessened settling time, and faster rise time highlight the effectiveness of the ZO-based FOPID controller in achieving optimal system response. Simulink is utilized to evaluate the performance of the system through simulation results. Additionally, the control algorithm is validated using an experimental Micro Lab Box in the laboratory.