基于自适应卡尔曼滤波的两电平背靠背系统准集中式直接模型预测控制

In conventional cascade-free control systems for two-level back-to-back power converters with PMSGs, the control structure is complex. Separate controls are needed for the front-end rectifier and the back-end inverter, and additional parameter tuning is required under different operating conditions. To address these challenges, a quasi-centralized direct model predictive control method using an adaptive Kalman filter (AKQC-DMPC) is proposed. This approach involves direct control of the DC-link voltage by a grid-side predictive controller, enabling centralized management of both the front-end rectifier and the back-end inverter. The adaptive Kalman filter is designed to correct system model inaccuracies and parameter uncertainties, thereby eliminating steady-state tracking errors in the DC-link voltage. Furthermore, to enhance the filter's resistance to disturbances, an adaptive rate function is introduced to dynamically adjust the fixed noise covariance matrix of the traditional Kalman filter. The effectiveness of AKQC-DMPC is demonstrated through experimental results. Its advantage is also established in comparison with traditional PI and QC methods, particularly in terms of robustness and dynamic performance.