Multiple-Voltage-Vector Model Predictive Control With Reduced Complexity for Multilevel Inverters
Yong Yang, Huiqing Wen, Mingdi Fan, Liqun He, Menxi Xie, Rong Chen, Margarita Norambuena, José Rodríguez
Abstract
Conventional model predictive control (MPC) suffers from unfixed switching frequency, heavy computational burden, and cumbersome weighting factors' tuning, especially for multilevel inverter applications due to a large number of voltage vectors. To address these concerns, this article proposes multiple-voltage-vector (MVV) MPC algorithms with reduced complexity and fixed switching frequency for T-type three-phase three-level inverters. First, MMVs are adopted during each control period, and their execution times are set according to the predefined cost functions. Second, weighting factors for balancing the neutral point (NP) voltage in the cost function are eliminated by utilizing redundant voltage vectors, which simplifies the control implementation. Third, through mapping the reference voltage in the first large sector, the calculation complexity for the execution times of voltage vectors in different large sectors becomes much lower. Finally, main experimental results were presented to validate the effectiveness of the proposed algorithms.