Capacitor Voltage Estimation of MMC Using a Discrete-Time Sliding Mode Observer Based on Discrete Model Approach

The modular multilevel converter (MMC) is widely used for high and medium voltage applications due to its several advantages over other multilevel topologies. The number of voltage levels that can be obtained at the ac output can be increased at the expense of a greater number of submodule (SM) capacitors. As the capacitor voltages need to be monitored for voltage balancing, the MMC system requires a large number of voltage sensors. This increases capital investment for the MMC system and failure points due to increased hardware interfaces. So far, a few works have been reported that uses observation techniques to either completely eliminate or reduce the number of sensors. This article presents a discrete-time sliding mode observer (DTSMO) for estimating the capacitor voltages and thus possibly eliminating all the voltage sensors of the SM capacitors. A switching function-based discrete mathematical model is used for the design framework of the proposed DTSMO. The errors in observation are taken as important performance indicators under both transient and steady-state conditions. It has also been observed that the proposed DTSMO delivers robust performance in presence of parametric uncertainty of capacitance values. The control algorithm is simulated in MATLAB/Simulink to validate the performance of the proposed observer. It is further validated through experimental results in a hardware prototype.