Design of MMC-STATCOM Controller Using an Adaptive PID Controller Supported by a Grey Model

Ali, Mohammed Moanes Ezzaldean
Kadhim, Qusay Shebeeb

Buletin Ilmiah Sarjana Teknik Elektro
Vol. 8 No. 1 (2026): February

The Static Synchronous Compensator (STATCOM) is recognized as one of the most advanced and effective technologies within the Flexible AC Transmission Systems (FACTS) family, due to its rapid dynamic response and high efficiency in regulating reactive power flow. However, conventional two- and three-level STATCOM topologies suffer from limited scalability and high harmonic distortion. This paper addresses these challenges by employing a STATCOM based on Modular Multilevel Converter (MMC). The significant contribution of this work is the introduction of a novel control strategy for MMC-STATCOM systems which is an adaptive PID controller integrated with a Grey Prediction Model. In the proposed scheme, the PID gains are continuously adapted based on predicted future error values obtained from the GM(1,1) grey model, rather than instantaneous measured errors, enabling proactive compensation under dynamic operating conditions. The performance of the proposed controller is evaluated in MATLAB/Simulink environment and by using a 12 MVA, 34.5 kV MMC-STATCOM system with a full-bridge topology consisting of 22 submodules per phase. Under balanced load condition, the results demonstrated that the adaptive grey-PID controller significantly reduced the total harmonic distortion (THD) of the grid current by 43.75% as compared to conventional PI controller. Under a severe unbalanced load condition, the total harmonic distortion of the grid current is reduced by 33.42%. Furthermore, the proposed controller successfully restored balance to the grid voltage and current and maintained a stable DC-link voltage under unbalanced load conditions. Additionally, the suggested controller achieved a fast-settling time of 0.04 s during transient conditions, this conclusively demonstrates its superior robustness and rapid dynamic response. Despite the additional computational effort introduced by the grey predictor model, it remains suitable for real time implementation due to its low order structure and limited data window.