Multi-level Model Predictive Controller with Satisfactory Optimization for Multi-Level Converters
Touati Mohamed Tamim,Shaoyuan Li,Jing Wu
DOI: https://doi.org/10.1016/j.simpat.2018.11.003
IF: 4.199
2018-01-01
Simulation Modelling Practice and Theory
Abstract:In this paper, a simplified Model Predictive Control (MPC) technique is proposed for a hybrid three-phase seven-level inverter based on Active Neutral Point Clamped (ANPC) converter cascaded to an H-bridge with a flying capacitor. This converter has the ability to balance switching losses, to withstand high voltages and to operate with fewer semiconductors. It is interesting for high-voltage and high-quality exigent applications. By introducing a trade-off relation between control actions reduction and the flying capacitor charge, the designed controller can maintain the capacitor voltage with less number of switching states. This reduces computations, eliminates common mode voltage, shortens optimization time, and improves output current quality. Moreover, in this work, a multilevel MPC architecture is presented, where the MPC approach is divided into two-level units. The major level unit is composed of separated sub-MPC controllers for each phase that track the three-phase current references, while the second level unit is composed of only one MPC model that balances the capacitor charge at its nominal values. An additional control design of the major level MPC is presented, where a satisfactory optimization of common mode voltage is used instead of weighting factor in the cost function, which simplifies the control approach, and solves the problem of selecting the right weighting factors. Simulation analysis shows the effectiveness of the proposed methods.