LIÊN KẾT WEBSITE
An Artificial Neural Network-Based Model Predictive Control Of Cascaded H-Bridge Multilevel Inverter
International Journal of Renewable Energy Research Số 3, năm 2022 (Tập 12, trang 1279-1288)
ISSN: 13090127
ISSN: 13090127
DOI:
Tài liệu thuộc danh mục:
Article
English
Từ khóa: Bridge circuits; Controllers; Electric inverters; Field programmable gate arrays (FPGA); Neural networks; Predictive control systems; Artificial neural network; Cascade H bridges; Cascade h-bridge; Cascaded H-bridge; Field programmables; Field-programmable gate array; Model predictive control; Model-predictive control; Multisteps; Programmable gate array; Model predictive control
Tóm tắt tiếng anh
In recent years, there has been an increasing interest in using Cascaded H-Bridge (CHB) for medium and high-power applications. Model predictive control (MPC) strategy has emerged as a promising option for controlling CHB with considerable advantages. However, the most significant disadvantage of MPC is the exponential increase of computational burden to solve the optimization, leading to an unacceptable amount of computing resources. Therefore, to overcome these difficulties, the ANNMPC approach for CHB is proposed in this paper. Firstly, the multistep MPC controller is designed and operated in simulation environment to generate the data required for training. Secondly, after being successfully trained, the neural network can be used to control the system without the need for MPC to avoid the heavy-duty computing problem. The performance of ANN-MPC is evaluated and compared to that of conventional multistep MPC. Finally, a FPGA based ANN-MPC controller employing the trained ANN is designed to control the experimental system with three-phase five-level CHB with LC filter and linear loads. Both simulation and experimental results verified the excellent control performance of the proposed ANN-MPC � 2022, International Journal of Renewable Energy Research.All Rights Reserved.
