Robust model predictive current control of grid-connected converter without alternating current voltage sensors

This study presents an improved hybrid parallel observer (HPO) based model predictive current control (MPCC) strategy to realise the source voltage sensorless control of grid-connected converters under inductance and resistance uncertainties. Since there is a little research on the sensorless control of MPCC-based converters, in this study, according to the linear extended-state observer theory and the adaptive filtering theory, a HPO which combines two different types of observers is constructed. With the new observer, the multivariable observation of system disturbance and model parameters is achieved. On this basis, through the combination of the HPO and the MPCC, a new HPO-MPCC control strategy is formed. The stability and parameter tuning of the HPO are analysed. To determine the initial value of the source voltage angle in the startup process of the converter, a new startup algorithm is proposed. A two-level grid-connected converter is built to verify the feasibility and effectiveness of the proposed strategy. The results show that, compared with the traditional algorithms, the new HPO has higher accuracy and stability; the HPO-MPCC achieves the source voltage sensorless control of the grid-connected converters and ensures the high precision and fast dynamic response characteristics of the conventional MPCC; the new startup strategy avoids the over-current problem of the conventional methods and improves the reliability of the whole system.