An adaptive backstepping observer for modular multilevel converter

This paper deals with the estimation problem in three-phase, three-level Modular Multilevel Converter (MMC) using nonlinear observer to realize an accurate estimation of unmeasured variables. It is shown that the desired states to be exponentially convergent if all system signals are bounded. An observer based on adaptive backstepping approach is used for capacitor voltages estimation from the arm currents measurement. The observer can track the capacitance value of module capacitor which is unknown as a parameter uncertainty with acceptable precision. In practical, it results attracting to evaluate the possibility of reducing the number of sensors, or even using these soft sensors as redundant for fault detection purposes. The nonlinear observer stability study by means of a Lyapunov theory guarantees the stability and convergence of the estimated quantities. The feasibility of the proposed solution is demonstrated by simulation results.