Coordinated Control of Static Reactive Power Generator and Fixed Capacitor Bank Under Reactive Power Prediction

To mitigate the issues of substandard power factor and degraded power quality in the grid caused by the high penetration of distributed photovoltaic integration, a coordinated control strategy with time gradient for Static Var Generator and Fixed Capacitor Bank based on a CNN-LSTM-Attention reactive power prediction model is proposed. To address the time delay problem in the existing coordinated control of Static Var Generator and Fixed Capacitor Bank, the load prediction model is trained using historical grid operation data, and combined with day-ahead reactive power demand prediction, a time gradient-based switching sequence diagram is established. This effectively avoids frequent switching of the capacitor banks, achieves reactive power optimization control in the grid, and meets the requirements for interest rate assessment and power quality.