基于MATLAB的模糊神经网络高压直流输电换流控制器的研究

在研究模糊逻辑控制技术、神经网络技术和高压直流输电（HVDC）系统的基础上，从原理上说明，对换流器两侧都采用模糊神经网络控制能有效地提高交流一直流（AC/DC）系统的动态特性和恒定性。整流侧和逆变侧分别提取直流线路电流、电压误差及其变化率作为模糊控制器的输入，输出作为神经网络的输入，分别控制电流和电压。用MATLAB对一典型12脉冲桥高压直流输电系统在传统控制和模糊神经控制下分别进行仿真。结果表明。与传统控制方法相比，当直流线路或者单相交流线路发生接地故障时。模糊神经控制能改善换流站直流电流和直流电压的恒定性，而且提高了交直流系统的暂态稳定性，并且双侧模糊逻辑控制要比单侧模糊逻辑控制效果好。

Study of HVDC fuzzy neural converter controller based on MATLAB

Based on the study of fuzzy logic control technology,neural network technique and high-voltage direct-current(HVDC)transmission system,it is explained theoretically that the adoption of fuzzy neural network control on both sides of converters caneffectively advance the dynamic behavior and steadiness of AC/DC system.The rectifier and the inverter correspondingly select thecurrent and voltage error of the DC line nearby their converter and the rate of change as input signal of fuzzy control,the output of fuzzypart as one of the input signals of neural network,so as to control current and voltage respectively.The MATLAB-based simulation fora typical HVDC 12-pulse bridge transmission system is carried out by traditional control method and fuzzy neural control methodrespectively.The simulation results show that compared with traditional control method,in case of grounding faults on DC lines orsingle-phase AC lines,fuzzy neural control can improve the steadiness of DC current and voltage in the converter substation as well asthe transient stability of AC/DC system.The simulation results also showthat the effect of fuzzy logic control on both sides of convertersis better than that on single side.