Abstract: Expected fault screening of smart distribution network is an important basis for system security situation assessment. In order to comprehensively and accurately perceive the security risk of smart distribution

network，a micro-synchronous phasor measurement unit（μPMU）with real-time，synchronicity，accuracy and

comprehensiveness of measurement data was introduced. A new method for the combination screening and sorting

of the expected fault combining based on high density sampling data weighting K-means clustering approach by

integrating intra-cluster and inter-cluster distances（KICIC）and cloud theory was proposed. Firstly，the failure

scenarios of each node of the smart distribution network were traversed，and the fault data sets were constructed.

Then，KICIC algorithm was used to conduct clustering analysis. Based on cloud digital features of cloud model，the

hazard of uncertainty of fault severity was quantitatively evaluated，and the expected fault set was outputed. Finally，the calculation results show that the expected fault screening method based on KICIC clustering and cloud model can reliably screen the high-risk expected fault sets in data mining level.

      Key words: smart distribution network；expected fault combination screening；weighting K-means clustering

approach by integrating intra-cluster and inter-cluster distances（KICIC）；cloud theory；micro-synchronous phasor

measurement uni（t μPMU）