基于VPMCD的中性点经低阻接地风电场集电线路单相接地智能区段定位

Intelligent Section Location for Single Phase Grounding Fault of Wind Farm Collection Line with Neutral Point Grounding via Resistor Based on VPMCD

针对风电场内多风机电源接入导致单相接地短路后定位精度差、效率低的难题,提出一种基于变量预测模型的类型判别(VPMCD)的集电线路智能区段定位新方法。首先,采集各测量装置的零序电流信息,考虑直接输入其波形序列存在过高的冗余和上传、运行成本,使用FFT提取零序电流的基频分量特征来构成定位样本库;然后,通过VPMCD方法建立各故障区段的变量预测模型,用于对测试故障样本进行分析;最后,应用上述的模型计算测试故障特征集合的误差平方和,并以其值最小为判定依据,自动、准确地定位故障区段。PSCAD/EMTDC实验仿真结果证明,新方法对多端电源系统有着良好的适应能力,且相较于支持向量机方法,无需优化调参,可实现复杂风电场网络的高精度、快速故障定位。

Aiming at the problem of poor locating accuracy and low efficiency after the single-phase grounding short-circuit caused by the access of multiple wind turbine sources in a wind farm,a new method for intelligent section locating of the collection line using variable predictive model based class discriminate(VPMCD)is proposed.Firstly,the zero sequence current information of each measuring device is collected.Considering the high redundancy,uploading and operating costs of directly inputting its waveform sequence,FFT is used to extract the fundamental frequency component features of zero sequence current to constitute a location sample library.Then,the variable predictive model of each fault section is established by VPMCD method,which is used to analyze the test fault samples.Finally,the above model is applied to calculate the sum of squares for error of the test fault feature set,and its minimum value is utilized as the judgment basis to automatically and accurately locate the fault section.PSCAD/EMTDC experimental simulation results prove that the new method has good adaptability to multi-terminal power systems.Compared with the support vector machine method,there is no need to optimize the parameter adjustment,which can realize the high precision and fast fault location of the complex wind farm network.