基于模型识别的有源配电网单相接地故障定位方法

分布式电源(distributed generation, DG)大量接入使配电网继电保护面临严峻挑战。单相接地故障发生概率大，但故障信息十分微弱，特别是在DG多样化故障输出特性的影响下，有源配电网单相故障定位的准确性常难以保证。为此，提出了一种基于模型识别的有源配电网单相接地故障定位方法。分析了配电网单相接地故障下DG的输出特性，建立了在不同故障位置下有源配电网的正序增广网络，构建了故障位置关于馈线出口以及DG输出电流的函数，建立了配电网单相接地故障位置模型，利用不同故障位置下短路电流矢量占比系数的差异性构建了新的故障定位判据。通过将单相接地故障定位问题转化为故障位置模型系数的求解问题，提高了故障定位的准确性和实用性。理论和仿真分析表明，在不同故障位置、不同过渡电阻下均能准确定位单相接地故障，具有原理清晰、故障特征量采集便捷、灵敏性和可靠性高的优点。

A fault location method for a single-phase fault for a distribution network with DGs based on model recognition

The large-scale integration of distributed power generation (DG) poses a serious challenge to the relay protection of distribution networks. The probability of single-phase ground fault occurrence is high, but the fault information is very weak, especially under the influence of the diverse fault output characteristics of DG. The accuracy of single-phase fault location in a distribution network with DGs is often difficult to guarantee. To this end, a model recognition-based single-phase grounding fault location method for a distribution network with DGs is proposed. The output characteristics of DG under single-phase to ground faults in the distribution network is analyzed, and a positive sequence augmented network of distribution network with DGs at different fault locations is established. A function of fault location on the feeder outlet and DG output current vector is constructed and a single-phase to ground fault location model in the distribution network is established. A new fault location criterion is constructed based on the difference in the proportion coefficient of a short-circuit current vector in different fault locations. The accuracy and practicality of fault location have been improved by transforming the problem of single-phase grounding fault location into the problem of solving the coefficients of the fault location model. Theoretical and simulation analysis shows that single-phase grounding faults can be accurately located at different fault locations and transition resistances, with the advantages of clear principles, convenient fault feature collection, and high sensitivity and reliability.