Privacy for the Smart Grid: Evaluating and enhancing Vehicle-to-Grid and Smart Metering approaches

Abstract

The Smart Grid is the envisioned next generation power grid. Key to a Smart Grid are its components with bidirectional communication capabilities such as Smart Meters and electric vehicles. Smart Meters enable a fine-grained monitoring and management of the energy flows from and to customer premises. Customers can use Smart Meter readings to learn their individual energy use patterns. Service providers such as grid operators, in turn, can use them to optimise the grid’s operation. The energy flows of electric vehicles connected to the grid can similarly be monitored and managed. The Vehicle-to-Grid concept leverages this and explores using electric vehicles and their batteries as a centrally coordinated, distributed grid resource. As such they can, e.g., temporarily store excess renewable energy available beyond the current demand. The Smart Grid’s monitoring capabilities are, however, also one of the key challenges to its introduction. While Smart Meter readings can reveal sensitive details about a customer’s behaviour and preferences, the information flows resulting from the Vehicle-to-Grid concept can reveal his whereabouts throughout a day.

In this thesis, we focus on these privacy issues. For the Vehicle-to-Grid concept, we initially consider an architecture built on current privacy-enhancing technologies and on anonymising Vehicle-to-Grid interactions. Based on this, we quantify the privacy achievable by such an approach and how information flows such as timing, location, and battery information can reduce privacy. Moreover, we outline potential approaches to protect customer privacy and evaluate them regarding their privacy benefits and effects on service functionality. For Smart Metering, we provide a detailed analysis of the privacy-preserving Smart Metering architecture mandated by German law that manages the information flows between a customer’s premises and remote service providers. This allows us to reveal possibilities of addressing shortcomings of the approach and other Smart Metering approaches that similarly rely on anonymising or pseudonymising Smart Meter data. Specifically, we put forward a service that allows to establish k-anonymity for Smart Metering data. Last but not least, by example of the German approach we discuss how to create an integrated privacy-preserving infrastructure that addresses both Smart Metering and Vehicle-to-Grid privacy. This approach offers two benefits. Firstly, it avoids costs for a separate Vehicle-to-Grid communication infrastructure. Secondly, the joint infrastructure significantly improves the level of Vehicle-to-Grid privacy that can be provided to customers.