Modified Q-Learning Method for Automatic Voltage Regulation in Wide-Area Multigeneration Systems

The state-estimation and optimal control of multigeneration systems are challenging for wide-area having numerous distributed automatic voltage regulators (AVR). This paper proposes a modified Q-learning method algorithm that aim to improve the convergence approach enhance dynamic response stability terminal multiple generators in experimental Western System Coordinating Council (WSCC) large-scale IEEE 39-bus test systems. testbed consists six-area, system ten connected AVRs. implementation shows promising results providing stable profiles other parameters across wide range AVR under different scenarios including N-1 contingency fault conditions. could provide significant improvement as compared conventional methods such using standalone and/or power stabilizers (PSS)