In view of different demands for power supply reliability of multi-type end users in different regions and the limitation of measuring user power supply reliability based on a single power supply reliability index at the distribution network side

a power supply demand division method for end users based on big data mining is proposed. Firstly

the differentiated power supply demand of endusers is quantified

and the information model of end-users’powersupply demand is constructed. Then

using integrated K-means and density- based spatial clustering of applications with noise big data mining to cluster end users

to achieve user classification. Finally

improved grey relational degree is used to divide the end user reliability level within the region. Through the simulation analysis of the power supply area with multi-type end users

combined with the analysis of comparative schemes

the validity of the proposed method based on big data mining is further verified.