Level-Increased Ishe Modulation Method for Modular Multilevel Converters Based on DDPG

Xinxiao Qin,Weihao Hu,Yubo Han,Yuanhong Tang,Jiachen Fan,Qi Huang,Hatem Khater,Frede Blaabjerg
DOI: https://doi.org/10.1109/tpel.2024.3486816
IF: 5.967
2024-01-01
IEEE Transactions on Power Electronics
Abstract:The output waveform quality of modular multilevel converters (MMCs) may not be good enough when there are only a few sub-modules (SMs). This paper proposes a technique to increase the quantity of equivalent output voltage levels of MMC by changing the capacitor voltage of certain sub-modules. This method significantly improves the performance of the MMC. In addition, the method combines a level-increased MMC with an improved selected harmonic elimination (iSHE) modulation, aiming to minimize the total harmonic distortion (THD) of the MMC output. The primary challenge of the iSHE modulation lies in solving a complex nonlinear mathematical model to determine the optimal switching angles. Deep reinforcement learning (DRL) has proven to be effective in addressing intricate multi-dimensional optimization problems within a continuous action space. Therefore, this paper utilizes the deep deterministic policy gradient (DDPG) algorithm to calculate the optimal switching angles of the MMC. The integration of DDPG algorithm enhances the operational performance of the MMC, thereby contributing to the progress of power electronics using AI.
What problem does this paper attempt to address?