风力发电系统变流器故障诊断

Fault diagnoses of converter in wind turbine system

针对风力发电系统运行过程中变流器一个或两个开关器件发生开路故障的情况,从逆变器PWM矢量控制出发,对不同的开路故障进行分析,得到故障情况下电流控制空间矢量图。考虑到故障情况下零点漂移、感应发电机的非线性性以及风力发电系统的强耦合性,对故障下的电流波形分析并提取故障特征,搭建以电流空间矢量特征函数作为输入的神经网络故障诊断模型,对逆变器21种可能的开路故障进行故障诊断和定位。仿真结果表明该方法运行速度快、诊断精度高。

As for open-circuit fault of one or two switch devices in operation, the PWM vector control strategy in each fault mode was analyzed and the space vector diagrams was drawn as well. The fault is featured by the zero drift during fault time, the nonlinear of induce generator and high-coupling of wind turbine system. In view of above reasons, the current plots were seriously considered and these features were extracted. So the input of neural network is the characteristic function of current space vectors. This method can diagnose and fix 21 kinds of probable open fault with fast processing speed and high preci-sion.