An Inductance Online Identification-Based Model Predictive Control Method for Grid-Connected Inverters With an Improved Phase-Locked Loop

Abstract

To simplify the tuning process and enhance the parameter robustness of the model predictive control (MPC) for grid-connected inverters, an improved phase-locked loop (PLL) and a new inductance online identification method are proposed in this article. First, an improved finite control set-based PLL is presented, which removes the PI controller, simplifying the tuning process. Next, a new inductance online identification method based on a model reference adaptive system is presented, which takes the actual grid voltage and the estimated one as the reference model and adjustable model, respectively. Here, a sliding mode observer (SMO) is proposed to estimate the grid voltage, and the inductance adaptive law is deduced by exploring the effects of the inductance on the estimated grid voltage. Furthermore, the sensitivity of the proposed SMO to frequency deviation is also analyzed, which shows that a proper sliding mode gain can be selected to reduce the effects of frequency deviation. Thus, a fixed grid frequency of 50 Hz can be used in the proposed SMO, indicating that it is not required to calculate the actual grid frequency using differential operation any more. Finally, experimental studies verify the validity of the proposed strategies.