Scholarlink

<p>This article introduces an advanced model predictive control (MPC) algorithm utilizing two vectors per sampling period for a two-level voltage source inverter (VSI) equipped with a low-pass LC filter. In contrast to conventional finite control set MPCs that rely on only one vector, the two-vector MPC yields superior steady-state performance and fixed frequency operation. The use of three vectors would incur greater computational expense and require additional switching operations. The proposed approach affords several benefits, including reduced low-order harmonics, improved steady-state ripple of the load voltage, and reduced computations. Additionally, fixed frequency operation eliminates resonances. Proposed method involves the utilization of two distinct objective functions and the optimization of selected switching states for the secondary objective function. By utilizing this approach, a comparable level of computational requirements to traditional MPC is achieved, while simultaneously maintaining superior steady-state performance. The proposed method is validated through the use of mathematical models, computational assessments, simulations, and experimental outcomes.</p>

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

Simplified two-vector predictive control for a VSI with LC filter for grid-forming applications

  • 期刊论文

  • 发表日期：

    2025年7月31日

查看全文
  • Siva Kumar Gannamraju,
  • Ravikumar Bhimasingu

摘要

This article introduces an advanced model predictive control (MPC) algorithm utilizing two vectors per sampling period for a two-level voltage source inverter (VSI) equipped with a low-pass LC filter. In contrast to conventional finite control set MPCs that rely on only one vector, the two-vector MPC yields superior steady-state performance and fixed frequency operation. The use of three vectors would incur greater computational expense and require additional switching operations. The proposed approach affords several benefits, including reduced low-order harmonics, improved steady-state ripple of the load voltage, and reduced computations. Additionally, fixed frequency operation eliminates resonances. Proposed method involves the utilization of two distinct objective functions and the optimization of selected switching states for the secondary objective function. By utilizing this approach, a comparable level of computational requirements to traditional MPC is achieved, while simultaneously maintaining superior steady-state performance. The proposed method is validated through the use of mathematical models, computational assessments, simulations, and experimental outcomes.