基于FOA-RBF神经网络的三相能馈型交流电子负载研究

Study of three-phase ac electronic load with energy-feedback based on FOA-RBF neural network

针对传统三相能馈型交流电子负载并网存在控制精度低、动态响应差、谐波含量高等问题,将基于RBF神经网络的PID控制用于电流控制器的设计。采用了果蝇优化算法对RBF神经网络的扩展参数进行了优化,提高了网络性能。针对并网逆变侧的控制策略进行了研究,并网逆变器采用双电流内环和带有滞环控制器的电压外环进行控制,可以提高并网电流质量并使能量单位功率因数回馈电网,达到提高能量利用率的效果;最后通过仿真验证了电流环的POA-RBF神经网络PID控制和电压环的滞环控制理论的正确性。

Considering the traditional control of AC electronic load with energy feedback with its low precision, poor dynamic and large harmonic content. PID control based on RBF neural network is applied into the current control loop. Fruit fly optimization algorithm was used to optimize the extended parameters of RBF neural network and improve the network performance. The control scheme of grid-connected inverter side has been designed. The grid-connected inverter is controlled by double current inner loop and voltage outer loop with hysteresis loop controller, energy feedback to power grid with unit power factor to save energy;Finally, the validity of POA-RBF neural network PID control of current loop and hysteresis control of voltage loop theory are verified through simulation.