Scholarlink

In this paper, an imitation learning-based model predictive control (IL-MPC) for hybrid active power filters (HAPF) is proposed. The proposed methodology ensures optimal control performance with a fixed switching frequency, significantly reducing the computational complexity typically associated with traditional model predictive control (MPC). First, the details the system description and mathematical modeling of a three-phase four-wire LC-HAPF. Then, a comprehensive design of the proposed IL-MPC with a neural network based nonlinear load predictive model is provided. Simulation results validate the effectiveness of the IL-MPC compared to finite control set MPC (FCS-MPC).

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Imitation Learning Based Model Predictive Control for Hybrid Active Power Filter

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Cheng Gong,
  • Chi-Seng Lam

摘要

In this paper, an imitation learning-based model predictive control (IL-MPC) for hybrid active power filters (HAPF) is proposed. The proposed methodology ensures optimal control performance with a fixed switching frequency, significantly reducing the computational complexity typically associated with traditional model predictive control (MPC). First, the details the system description and mathematical modeling of a three-phase four-wire LC-HAPF. Then, a comprehensive design of the proposed IL-MPC with a neural network based nonlinear load predictive model is provided. Simulation results validate the effectiveness of the IL-MPC compared to finite control set MPC (FCS-MPC).