Model-Free Predictive Current Control of a Modular Multilevel Converter Based on Nearest-level Modulation
- Resource Type
- Conference
- Authors
- Zhao, Fangyuan; Zhang, Yongchang
- Source
- 2023 26th International Conference on Electrical Machines and Systems (ICEMS) Electrical Machines and Systems (ICEMS), 2023 26th International Conference on. :4537-4542 Nov, 2023
- Subject
- Aerospace
Components, Circuits, Devices and Systems
Power, Energy and Industry Applications
Robotics and Control Systems
Transportation
Current control
Multilevel converters
Inductance
Modulation
Predictive models
Power system harmonics
Mathematical models
Predictive control
ultra-local model
modular multilevel converter
robustness
- Language
- ISSN
- 2642-5513
The efficacy of model-based predictive current control (MBPCC) for grid-tied modular multilevel converters (MMC) is contingent upon the precision of the system model. Any fluctuations in the system parameters can lead to a decline in the control performance of power and grid currents. To address this challenge, the proposed method advocates an ultra-local model that is continuously updated on-time based on the voltages and currents from previous control periods, as opposed to the conventional accurate model of MMC. The ultra-local model is utilized to calculate voltage reference for the purpose of nullifying current errors, employing the principle of deadbeat control and synthesizing them through NLM. To validate the effectiveness of the proposed method compared to MBPCC, a comparison is conducted under AC equivalent inductance mismatch conditions. MATLAB/Simulink results confirm the superior performance of the proposed method.