基于数据驱动建模与免矩阵求逆模型预测控制的多激励单元无线功率传输系统动态性能增强

The multiexcitation-units wireless power transfer (MEU-WPT) system supports multilevel power applications by adjusting the number of excitation units. Its inherent redundancy enhances reliability against inverter failures, making it ideal for wireless electric vehicle and underwater vehicle charging. However, the incorporation of the MEU structure increases the number of control variables (such as the phase-shift angles of the inverters) and introduces cross couplings among different power units, resulting in complex power transfer channels and interconnected feedback loops, where variations in one control variable generate disturbances in other channels. Consequently, achieving high dynamic performance in output regulation becomes challenging without effective coordination of all control variables. This article proposes a constrained model predictive control (MPC) method aimed at achieving high dynamic performance in output regulation for the MEU-WPT system with cross couplings. Initially, a data-driven simplified refined instrumental variable method is employed to model the MEU-WPT system, capturing the dynamics associated with the cross couplings. Subsequently, a Laguerre-network based multivariable MPC control law is derived from the data-driven models. A matrix-inversion-free quadratic programming solver is developed to enable computationally-efficient constrained optimization. Furthermore, an event-triggered mechanism is integrated into the control framework, demonstrating the potential for rapid switching between different input power allocations among excitation units. An MEU-WPT prototype featuring two excitation units is constructed to validate the feasibility of the proposed control approach.