Model Predictive Control for Multi-Level Power Converters
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional power converter control methods, such as PI control, face challenges in efficiently managing multiple control parameters and maintaining closed-loop performance due to computational demands and switch timing constraints, especially in advanced control algorithms for multi-level power converters.
Innovation Solution
The implementation of model predictive control (MPC) using a plurality of switching matrices to select and adjust the switching states of power converter switches based on a multi-objective function, optimizing switch timing and minimizing errors, switching frequency, and DC bus voltage balancing.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If model predictive control is implemented to optimize multiple control parameters, then control performance and efficiency are improved, but computational complexity and implementation difficulty increase
Solution Approach 1:
The control algorithm is segmented into modular components: state prediction module, cost function evaluation module, and switching matrix selection module. Each module handles specific computational tasks independently, reducing overall complexity while maintaining optimization capability across multiple control parameters
Solution Approach 2:
Switching matrices are pre-calculated and stored based on predicted future states of the power converter. This preliminary action allows the controller to quickly select optimal switching sequences without real-time computational burden, improving control performance while reducing implementation complexity
2Reliability
If advanced control algorithms are used to monitor multiple control parameters simultaneously, then closed-loop performance is improved, but computational demands increase
Solution Approach 1:
The most computationally intensive calculations are extracted from the real-time control loop and performed offline or in advance. Switching matrices and optimal control sequences are pre-determined based on system models, reducing real-time computational demands while maintaining reliable closed-loop performance through multi-parameter monitoring
3Loss of energy
If switching frequency is reduced to minimize losses, then energy efficiency is improved, but current tracking quality deteriorates
Solution Approach 1:
The switching frequency is made dynamic rather than fixed. The controller adapts switching frequency based on operating conditions, load requirements, and current tracking needs. This allows the system to minimize switching losses during light loads while maintaining high current tracking quality during dynamic transitions or heavy loads
Data Source
AI summary
Embodiments are directed to a model predictive control for power electronics. The model predictive control includes a plurality of switching matrices defining potential states of a plurality of power converter switches of a multi-level power converter and a control. The control is configured to select a current switching matrix from the switching matrices that models the multi-level power converter in a current state. The control determines a targeted switching matrix from the switching matrices that best aligns with a targeted state based on alignment with a multi-objective function and changes with respect to the current state. The control adjusts a switch state of the power converter switches based on the targeted switching matrix. The control sets the current switching matrix to the targeted switching matrix and monitoring for changes with respect to the multi-objective function and the current state.