Hybrid Model Predictive Current and Voltage Control for LCL-Filtered Grid-Connected Inverter

The inductive-capacitive-inductive (LCL)-filtered grid-connected inverter features a high-order plant, complex parameter design, and vulnerability under disturbance factors. The transfer-function-based active damping (TFAD) scheme is commonly adopted to mitigate the resonance problem but with cascaded control loops and high design complexity. On the contrary, multiobjective optimization can be easily realized in model predictive control (MPC) via a predefined cost function. As the coupling dynamic among state variables, a hybrid cost function, which contains the grid current, inverter-side current, and capacitor voltage tracking, is designed to guarantee system stability. The optimized control law is generated according to the minimization of the designed cost function, which is further transformed into the driving pulse by the modulator stage. In this process, the weighting factors are exactly arranged to place the desired closed-loop poles. The control law can ensure zero error tracking of grid current reference, which is better than the state feedback way. The hybrid cost function with modified reference is also proposed to reject the harmonics caused by the unhealthy grid condition. The control delay is compensated by using the forecast reference and the predictive state variable estimated from the observer. The proposed scheme performs the intuitive concept, reduced tuning complexity, and fast dynamic response. Simulated results and experiments are presented to verify the proposed control strategy.