The ameliorations in high-precision phasor measurement units (μPMUs) and synchrophasor units have accommodated the distribution grid with peculiar visibility. Therefore, investigating the challenges of uncertainty consideration on precise fault detection in microgrids has become a new research milestone. This paper presents an effective data-driven stochastic method that justifies the adoption of only two μPMUs that are communicating under an IoT-based umbrella to detect and allocate irregularities in a microgrid. The proposed method has the ability to operate under a variety of case studies and scenarios including but not limited to the capacitor bank switching, distributed energy resources (DERs) diversity and high impedance fault occurrence, whilst considering the uncertainty in load, without installing individual sensors. Furthermore, a two-point estimate approach is utilized to model the uncertainties of the problem. Not only does the proposed stochastic framework benefit from the voltage magnitude measurement, but it also utilizes its angle in event allocation, which manifests better performance compared to ordinary voltage and current sensors. The simulation results on the proposed microgrid indicate the high accuracy and a sound success is obtained under a variety of case studies. The results show the high accuracy and applicable aspect of the proposed data-driven approach for fault allocation using a few μPMUs in the IoT context.

高精度相量测量单元（μPMU）和同步相量单元的改进使配电网具有了独特的可视性。因此，研究不确定性因素对微电网中精确故障检测的挑战已成为一个新的研究里程碑。本文提出了一种有效的数据驱动的随机方法，该方法证明仅采用两个在基于物联网的保护下进行通信以检测和分配微电网中的异常情况的μPMU是合理的。所提出的方法具有在各种案例研究和场景下运行的能力，包括但不限于电容器组切换，分布式能源（DER）分集和高阻抗故障发生，同时考虑了负载的不确定性，而无需安装单个传感器。此外，采用两点估计方法对问题的不确定性进行建模。所提出的随机框架不仅受益于电压幅度测量，而且还利用其角度进行事件分配，与普通的电压和电流传感器相比，它表现出更好的性能。所提出的微电网的仿真结果表明，该方法具有很高的准确性，并且在各种案例研究中均取得了成功。结果表明，在物联网环境中，使用几个μPMU进行故障分配的数据驱动方法具有较高的准确性和适用性。与普通的电压和电流传感器相比，具有更好的性能。所提出的微电网的仿真结果表明，该方法具有很高的准确性，并且在各种案例研究中均取得了成功。结果表明，在物联网环境中，使用几个μPMU进行故障分配的数据驱动方法具有较高的准确性和适用性。与普通的电压和电流传感器相比，具有更好的性能。所提出的微电网的仿真结果表明，该方法具有很高的准确性，并且在各种案例研究中均取得了成功。结果表明，在物联网环境中，使用几个μPMU进行故障分配的数据驱动方法具有较高的准确性和适用性。