Power Losses Reduction of T-Type Grid-Connected Converters Based on Tolerant Sequential Model Predictive Control

Abstract

Three-level T-type converters (3LT <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup> C) with inductance–capacitance–inductance filters have high power quality of grid current, particularly in the low-voltage system. However, the switching loss and conduction losses degrade the overall efficiency of 3LT <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup> C. The generated heat due to the losses endangers the reliability of the device and shortens its service life. To relieve the power losses as well as maintain high-quality current output, a low-loss-tolerant sequential model predictive control (LL-TSMPC) is proposed. First, based on the energy loss analysis under switching transitions per commutation, the prediction models of both switching loss and conduction loss are established. Second, the total loss reduction is considered as one control objective in LL-TSMPC. Finally, because multiobjective model predictive control (MPC) requires selecting appropriate weighting factors, complicating the design to solve this problem, TSMPC-based lexicographic optimization is introduced to eliminate the tradeoff factors and simplify the MPC controller design. The proposed method is tested on the hardware platform with the rated power of a 10-kW prototype. The experimental results show that the LL-TSMPC method can effectively reduce the switching losses while maintaining the high power quality of the grid current.