HVDC中单神经元自适应PID控制方法研究

介绍了将具有自学功能的单神经元模型与常规的比例、积分、微分(PID)控制方法相结合,设计的基于单神经元自适应PID控制器。其控制规律是在高压直流输电(HVDC)系统以整流侧定电流的PID为研究对象,将传感器测得的实际直流电流与参考值比较,再将两者误差信号反馈给PID控制器,产生延迟相角信号。依据误差信号的变化调节延迟相角,从而保持直流线路的恒定。所设计的控制器由实际直流电流与电路参考值之差输入到状态变换器获得的3个量作为神经元的输入,并基于Hebb学习规则、联想式学习策略推导出单神经元的权值调整规律和参数整定方法。采用Simulink典型6脉冲桥HVDC系统的整流侧定电流控制模型仿真,实例说明了神经元算法的实现,其设计的控制器效果好。

Application of adaptive PID control based on single neuron in HVDC

An adaptive PID(Proportional-Integral-Differential) controller based on single neuron is designed,which combines the single neuron model with self learning function and the regular PID control.To take the fixed current control of HVDC(High Voltage Direct Current transmission) as an object investigated,the control rule is to compare the reference current with the actual current measured by the sensor and to send the difference between them to PID controller as a feedback for generating the delay phase angle.The delay phase angle is adjusted according to the change of the difference signal to keep the current of DC line steady.When the difference signal is put into the state transformer,three outputs are produced,which are the inputs of single neuron.Based on Hebb learning rule and associate learning strategy,the adaptation law for adjusting weights of the single neuron and the way for setting parameters are deduced.With the typical six-pulse bridge HVDC system of Simulink,the single neuron arithmetic is simulated and results show its effectiveness.