Outsourced and Privacy-Preserving K-means Clustering Scheme for Smart Grid

Currently, the k-means clustering algorithm is generally used to mine the available characteristics from the massive power consumption data, so as to provide high-quality and customized electricity services for grid users. However, these data is sensitive and can be used to speculate on large amounts of private information, such as users living habits. To address these problems, this paper proposes an outsourced and privacy-preserving k-means clustering scheme (OPKM). Firstly, the additive secret sharing technology is used to split user data into two shares, which are sent to two cloud servers. Secondly, secure distance, multiplexer, minimum and division protocols are designed to achieve the secure cluster initialization and secure clustering for k-means algorithm. The experimental results with real electricity dataset show that the clustering accuracy and efficiency of the proposed OPKM scheme is better compared the existing works.

I. Introduction

With the continuous development of big data, the traditional grid has been unable to satisfy the electricity requirements of efficiency, intelligence and security in the current society. As an emerging grid mechanism, smart grid integrates information technology, grid infrastructure and control technology, and basically realizes the automation, intelligence and informatization of the grid. On the other hand, with the increase of the grid energy demand and the expansion of the grid scale, the grid client generates a steady stream of electricity consumption data. How to effectively analyze the potentially valuable information from these massive data has become a research hotspot in the current smart grid.