Adaptive hybrid PSO-GD optimized phase-locked loop for robust grid synchronization in renewable energy systems

• Published in: Compel Vol. 44; no. 4; pp. 566 - 596
• Main Authors: Rajak, Mrinal Kanti, Pudur, Rajen
• Format: Journal Article
• Language: English
• Published: Bradford Emerald Publishing Limited 03.07.2025; Emerald Group Publishing Limited
• Subjects: Accuracy; Adaptation; Adaptive algorithms; Algorithms; Alternative energy sources; Critical components; Disturbances; Dynamic response; Energy resources; Filtration; Frequency variation; Fuzzy logic; Harmonic distortion; Machine learning; Optimization techniques; Parameter estimation; Parameters; Particle swarm optimization; Phase error; Phase locked loops; Power transfer; Real time; Renewable energy; Renewable resources; Stability analysis; State space models; Synchronism; Tuning; Voltage sags; Adaptive PLL; Dynamic parameter tuning; Gradient descent; Renewable energy; Power quality; Particle swarm optimization; Grid synchronization
• ISSN: 0332-1649; 2054-5606; 2054-5606
• DOI: 10.1108/COMPEL-11-2024-0463

Purpose This paper aims to address the challenges of maintaining accurate grid synchronization in renewable energy systems under dynamic grid disturbances. Grid-connected renewable energy systems require precise synchronization with the utility grid to ensure stable power transfer and maintain power quality. Phase-locked loops (PLLs) serve as critical components by accurately tracking the grid’s phase angle, frequency and voltage parameters in real time. However, the increasing penetration of renewable sources introduces significant grid disturbances including voltage sags, phase jumps, frequency variations and harmonic distortions that challenge conventional PLL performance. Under these dynamic conditions, maintaining accurate grid synchronization becomes essential for preventing system instability, ensuring power quality and meeting grid code requirements. Design/methodology/approach A novel hybrid PLL method is proposed, combining particle swarm optimization (PSO) and gradient descent (GD) algorithms. The adaptive model continuously tunes PLL parameters, enabling real-time responsiveness to grid fluctuations, including voltage sags, phase shifts and harmonic distortions. The mathematical framework incorporates state-space modeling and Lyapunov stability analysis. Advanced loop filter architectures combining multiple delayed signal cancellation (MDSC) with complex coefficient filtering (CCF) provide enhanced harmonic rejection. Real-time parameter adaptation ensures optimal performance under varying grid conditions. Findings The hybrid PSO-GD PLL demonstrates superior performance over traditional synchronization techniques (SRF-PLL, DDSRF-PLL, MSOGI-PLL) across multiple metrics. In steady-state operation, it achieves frequency error of ±0.1 Hz and phase error of ±0.5° compared to ±0.5 Hz and ±2.0° for SRF-PLL. Dynamic response shows significant improvements with 25 ms settling time and ±0.3 Hz frequency overshoot, versus 40 ms and ±2.0 Hz in SRF-PLL. The system exhibits enhanced stability with 65° phase margin, 120 Hz bandwidth and excellent performance under grid fault conditions through auto-tuning capability. Originality/value This work introduces an adaptive, hybrid optimization framework for PLL, specifically enhancing grid-tied renewable systems’ synchronization accuracy and resilience. Integrating PSO’s global search capabilities with GD’s local refinement creates a unique dual-stage optimization process that surpasses traditional single-algorithm approaches. The mathematical framework’s incorporation of advanced filtering techniques (MDSC and CCF) with Lyapunov-based stability analysis represents a novel approach to PLL design. Moreover, the auto-tuning capability through real-time parameter adaptation addresses the critical challenge of maintaining synchronization in increasingly complex grid environments with high renewable penetration.