无线电能传输系统输出电压控制与效率优化

Outputvoltage control and efficiency optimization of wireless power transfer system

关键词:  无线电能传输  神经网络控制  移相控制  扰动观测法  效率优化

Key words:  wireless power transfer  neural network control  phase-shift control  disturbance observation method  efficiency optimization

摘要:

为解决磁场耦合型无线电能传输(WPT)系统在实际应用中因负载动态变化、耦合系数波动、元件参数漂移所导致的输出电压不稳定,以及传输效率相对较低的问题,提出了一种基于径向基函数神经网络(RBFNN)和改进的扰动观测(P&O)算法的双边移相协同控制策略。首先,构建并分析了双边LCC补偿拓扑的WPT系统数学模型,针对其模型的非线性和不确定性,设计了一种具有在线自学习能力的RBFNN控制器。通过实时采集系统输出误差信息动态调整网络权值,直接生成精准的移相角控制信号,通过控制接收端可控整流电路移相角来调节系统输出电压。有效克服了传统控制依赖于精确模型、适应性差的缺点。其次,为了在保证输出电压稳定的同时最大化系统的传输效率,发射端采用变步长的P&O算法动态调节逆变电路移相角实现系统的最大效率跟踪。最后,搭建了一台实验样机进行验证。结果表明:系统输出电压呈现良好的动态响应性能,可实现期望电压的无超调跟踪；且在扰动作用下表现出强鲁棒性,电压波动范围小于1%；同时系统效率最大提升14.9%,充分证明了所提方法的有效性。

Abstract:

To address the issues of unstable output voltage and relatively low transfer efficiency in magnetically coupled wireless power transfer (WPT) systems caused by dynamic load variations, coupling coefficient fluctuations, and component parameter drift in practical applications, this paper proposed a dual-side phase-shift cooperative control strategy based on a radial basis function neural network (RBFNN) and an improved perturb and observe (P&O) algorithm. First, this paper constructed and analyzed a mathematical model of the WPT system with a bilateral LCC compensation topology. To address the nonlinearity and uncertainties of the model, this paper designed an RBFNN controller with online self-learning capability. By collecting real-time system output error information to dynamically adjust network weights, it directly generated accurate phase-shift angle control signals and regulated the system output voltage by controlling the phase-shift angle of the controllable rectifier circuit at the receiving end. This effectively overcame the shortcomings of traditional control methods relying on precise models and poor adaptability. Second, to maximize the transfer efficiency while ensuring stable output voltage, the transmitting end adopted a variable-step P&O algorithm to dynamically adjust the phase-shift angle of the inverter circuit for maximum efficiency tracking. Finally, this paper built an experimental prototype for verification. The results demonstrate that the system output voltage exhibits excellent dynamic response performance and can achieve overshoot-free tracking of the desired voltage; it shows strong robustness under disturbances, with a voltage fluctuation range of less than 1%; meanwhile, the system efficiency increases by a maximum of 14.9%, which fully proves the effectiveness of the proposed method.