Reactive Power Minimization Modulation Strategy for NPC-SRDAB based on Particle Swarm Optimization

This paper presents a minimization modulation strategy for reactive power in neutral point clamped series resonant dual active bridge (NPC-SRDAB) with four degrees of freedom (DOF), utilizing the particle swarm optimization (PSO) algorithm. The objective is to enhance the efficiency of NPC-SRDAB across the entire power range. By considering the external characteristics of the high-frequency output waveforms, a Fourier series-based high frequency link mathematical model is established for AC output voltage, current, transmitted active power, and reactive power. The PSO algorithm is employed for offline computation, determining the optimal combination of DOF that minimizes reactive power for different voltage gains and power levels. To represent the behavior of each DOF across the entire power range and a wide range of voltage gains, shallow artificial neural network (SANN) is employed for fitting purposes. The resulting fitted function can be utilized in online digital control.