Data Mining and Privacy of Personal Behaviour Types in Smart Grid

Privacy protection is one of the key requirements of smart grids. To understand the importance of privacy threats it is necessary to study nature of power signals. In this paper, we propose a well-known statistical method which relies on the empirical probability distribution. The method is used to reveal trends in the power signal data and how these trends are changed if a) different data sampling rates are assumed, and b) a privacy algorithm is applied to protect the power data of different home appliances. Our results suggest that the privacy of personal behaviour types is exposed even if relatively infrequent measurements are obtained. On the other hand, battery-assisted home energy management solutions are more likely to protect the customers.