Soft Computing-Based Damping Controllers With Online Parameters Tuning for Stability Enhancement of Power Systems

Power system stability continues to be a major challenge as modern grids grow more complex, uncertain, and increasingly reliant on renewable energy sources. This paper presents two new Neuro-Fuzzy controllers for Static Synchronous Compensators (STATCOMs): the Direct Chebyshev Wavelet-Based Neuro-Fuzzy Controller (DNF-CW), which adapts parameters online using fixed-structure rules, and the Indirect Chebyshev Wavelet-Based Neuro-Fuzzy Controller (IDNF-CW), which uses an online identifier to measure plant sensitivity. Chebyshev wavelet-based neural networks are utilized in the consequent part of both controllers to enable accurate local modeling and improved damping performance. The proposed methods are evaluated using a Single Machine Infinite Bus (SMIB) system and the IEEE 9-bus multi-machine system under a variety of fault and loading conditions. Benchmark comparisons include a conventional Indirect Adaptive Neuro-Fuzzy Takagi–Sugeno–Kang (IDNF-TSK) based controller, a configuration without STATCOM (No STATCOM), and a configuration with STATCOM but without auxiliary control (No Control). In the SMIB scenario, the IDNF-CW achieves a 40% reduction in settling time compared to the IDNF-TSK. In the more demanding multi-machine setup, the IDNF-CW restores stability within 3 seconds after a sequence of faults, outperforming DNF-CW and IDNF-TSK. Additionally, reductions of over 53% in the Integral of Time-weighted Absolute Error (ITAE) and 36% in the Integral of Absolute Error (IAE) are observed. These tests under multiple fault conditions and 10% measurement noise confirm stable operation, with overshoot limited to 3.57%–4.7% and minimal impact on settling time. These findings highlight the effectiveness of combining Chebyshev wavelets, adaptive control, and indirect architectures for enhancing power system stability.