DESIGN OF AN ITERATIVE METHOD INTEGRATING BOILER DYNAMICS AND DEEP Q LEARNING WITH FUZZY LOGIC AUGMENTATION FOR AGC BASED WIND POWER SYSTEM OPTIMIZATION

The need for enhanced efficiency, stability, and reliability in modern power systems has become increasingly critical with the escalating integration of renewable energy sources. Existing methodologies often fall short in balancing the intricate trade-offs among these factors, predominantly due to their inadequate adaptability to dynamic operating conditions and Automatic generation Control (AGC) based renewable integration systems. This paper introduces a novel approach by integrating Boiler Dynamics with Deep Q Learning (DQL) and augmenting it with Fuzzy Logic and virtual inertia to optimize power system operations. The Boiler Dynamics component is crucial for understanding the thermal aspects of power generation, which, in combination with DQL, facilitates autonomous, adaptive decision-making. The incorporation of virtual inertia provides an additional layer of stability by mimicking the inertial response of traditional synchronous generators, thereby enhancing the system's ability to handle frequency fluctuations. The proposed model distinctly outperforms conventional systems by achieving an overall system efficiency of 90-95%, while maintaining frequency deviations within ±0.05 Hz, thus ensuring grid stability levels. Control errors are minimized to less than 1%, reflecting the system’s capability to maintain near-optimal conditions. Additionally, the model supports substantial renewable integration, achieving renewable penetration levels of 30-40%, thereby enhancing the sustainability of power generation.