基于C-DWT和FFNN的风电场接地故障诊断方法

GROUND FAULT DIAGNOSIS METHOD FOR WIND FARM BASED ON C-DWT AND FFNN

为及时、准确地检测风电场交流输电线路的接地故障并对其进行分类,同时解决故障电阻、故障起始时刻和故障位置给故障诊断精度带来的影响,提出一种基于Clarke变换、离散小波变换(DWT)和前馈神经网络(FFNN)相结合的风电机组交流输电线路故障检测及分类方法。该方法通过Clarke变换对测量的三相电压信号进行分解生成γ分量、α分量和β分量,采用DWT就3个分量进行高频分量的提取,将5种统计学方法应用于高频分量D1,生成最终故障特征值。将γ分量的故障特征值和γ分量、α分量和β分量的故障特征值分别与FFNN相结合以实现精准的风电机组接地故障检测和分类。通过1400个故障案例验证该方法可不受故障电阻、故障起始时刻和故障位置的影响,有效对风电场交流输电线路进行故障诊断。

In order to timely and accurately detect grounding faults in wind farm AC transmission lines and classify them.At the same time,to solve the influence of fault resistance,fault starting time and fault location on fault diagnosis accuracy,a fault detection and classification method for wind turbine AC transmission lines based on the combination of Clarke transform,discrete wavelet transform(DWT)and feedforward neural network(FFNN)is proposed.This method decomposes and generates the measured three-phase voltage signal through Clarke transformγcomponent,αcomponent andβcomponent,DWT is used to extract high-frequency components from three components,and five statistical methods are applied to high-frequency component D1to generate the final fault feature values.Takeγcomponent of fault eigenvalues andγComponent,αcomponent andβcomponent of fault eigenvalues are combined with FFNN to achieve accurate detection and classification of wind turbine grounding faults.Through 1400 fault cases,it has been verified that this method can effectively diagnose faults in wind farm AC transmission lines without being affected by fault resistance,fault initiation time,and fault location.