Intelligent Power Feedback Control for Motor-Generator Pairs: A Machine Learning-Based Approach

The escalating integration of renewable energy sources poses notable challenges to the stability of power systems, encompassing diminished inertia, frequency instability, and issues related to grid resilience. Proposals have been made for a motor-generator combination system as a practical method to confront these issues by supplying mechanical inertia and unlinking renewable energy sources from the grid. Nevertheless, current control methodologies, including source-grid phase difference control, demonstrate a pronounced sensitivity to frequency fluctuations, thereby constraining their practical application. This paper introduces an advanced control methodology for MGP systems through the incorporation of an optimized power feedback control strategy alongside intelligent tuning mechanisms. The proposed strategy utilizes machine learning-based predictive control and adaptive proportional-integral (PI) tuning to dynamically modify the source-grid phase difference, thereby ensuring stable active power transmission in the face of grid disturbances. Furthermore, a hybrid renewable energy integration framework is proposed, which amalgamates MGP with energy storage systems to enhance reliability. Simulation and experimental validation reveal that the proposed methodology significantly mitigates frequency sensitivity, bolsters grid stability, and preserves active power transmission within established limits. A comparative analysis with traditional control techniques underscores the efficacy of the enhanced system in practical applications. The outcomes of this investigation furnish a solid foundation for the implementation of MGP systems within high-penetration renewable energy grids, thereby facilitating the development of more resilient and efficient power networks.