基于GA-BP神经网络的双馈风力发电机故障诊断

Fault Diagnosis of Doubly-fed Wind Turbine Based on GA-BP Neural Network

双馈风力发电机在风力发电系统中得到了广泛应用,但由于其工作条件及自身结构原因,导致故障发生率较高。BP神经网络作为一种多层前馈网络,在电机故障诊断分析领域应用成熟。为避免陷入极小值问题,利用GA遗传算法对其优化,建立GA-BP神经网络模型,对双馈风力发电机定子匝间短路特征进行分析,以定子电流为故障信号,经快速傅里叶分解得到的电流特征量作为样本输入,输出为预期的故障类型,进而实现不同程度的定子匝间短路故障的诊断与识别。

Doubly-fed wind turbine has been widely used in wind power generation system.However,due to its working conditions and its own structure,the failure rate is high.As a multilayer feedforward network,BP neural network is widely used in the field of motor fault diagnosis and analysis.In order to avoid falling into the minimum problem,the GA genetic algorithm is used to optimize it.A GA-BP neural network model is established to analyze the inter-turn short-circuit characteristics of the stator of the doubly-fed wind turbine.Taking the stator current as the fault signal,the current characteristic quantity obtained by fast Fourier decomposition is used as the sample input,and the output is the expected fault type,so as to realize the diagnosis and identification of stator turn to turn short circuit fault in different degrees.