Improved Model Predictive Control of Grid‐Connected Inverters Based on Online Inductance Identification for Power Quality Improvement

ABSTRACT

The growing penetration of renewable energy sources into the power grid has raised concerns about power quality, as fluctuating filter parameters often lead to control instability and harmonic distortion in grid‐connected inverters. To reduce the influence of inductance‐identification errors caused by grid‐frequency deviations, an MPC strategy based on improved online inductance identification is developed. First, an improved grid‐voltage estimation method combining a sliding‐mode observer and two cascaded quasi‐resonant filters is developed. In the model reference adaptive system structure, the measured grid voltage is used as the reference model, while the grid voltage estimated from the measured current is used as the adjustable model. The quasi‐resonant filters remove the DC component in the estimated voltage caused by the current‐sampling offset and suppress the high‐frequency chattering of the observer. Since the resonant frequency is fixed at the nominal grid frequency, a compensation stage is further introduced to compensate for the dominant amplitude and phase errors introduced by the quasi‐resonant filters caused by grid‐frequency deviations. The identified inductance is subsequently applied to the MPC prediction model. Hardware‐in‐the‐loop results show that the proposed method limits the absolute mean inductance‐identification error to approximately 0.2 mH at 49 Hz and 51 Hz, reduces grid‐current harmonic distortion, and maintains satisfactory voltage and current tracking under the considered operating conditions.