Robustness improvement model predictive control strategy based on Luenberger observer for Y-type modular multilevel converter

The Y-type modular multilevel converter (Y-MMC) has high reliability due to its high modularity. It can promise AC/AC conversion for high-voltage and large-capacity fractional frequency transmission systems (FFTS). The Y-MMC under linear control has a poor ability to adapt to changing working conditions, resulting in unstable transmission, slow response, and large harmonics and overshoot. The Robustness Improvement Model Predictive Control strategy (RI-MPC) for Y-MMC under nonideal conditions is presented in this paper to enhance robustness under unbalanced conditions and adaptability to load changes. It points out that the conventional MPC has low robustness when Y-MMC encounters parameters difference due to faulty line, designs an improved outer-loop PI control to enhance robustness to provide an excellent reference value for the inner-loop current, combines the Luenberger observer with MPC and limits the parameters with Jury criterion to improve stability, and uses a two-step prediction to compensate for delay and reduce harmonics. Through software simulation and hardware experiments, it is proved that compared with the traditional PI and MPC method, the RI-MPC method is more robust, faster, and has application value.