Academic and research departmentsInstitute for Communication Systems, School of Computer Science and Electronic Engineering.
Integrating information and communication technologies into the power generation, transmission and distribution system provides a new concept called Smart Grid (SG). The wide variety of devices connected to the SG communication infrastructure generates heterogeneous data with different Quality of Service (QoS) requirements and communication technologies. An intrusion Detection System (IDS) is a surveillance system monitoring the traffic flow over the network, seeking any abnormal behaviour to detect possible intrusions or attacks against the SG system. Distributed fashion of power and data in SG leads to an increase in the complexity of analysing the QoS and user requirements. Thus, we require a Big Data-aware distributed IDS dealing with the malicious behaviour of the network. Motivated by this, we design a distributed IDS dealing with anomaly big data and impose the proper defence algorithm to alert the SG.This paper proposes a new smart meter (SM) architecture,including a distributed IDS model (SM-IDS). Secondly, we implement SM-IDS using supervised ML algorithms. Finally, a distributed IDS model is introduced using federated learning.Numerical results approve that Neighbourhood Area Network IDS (NAN-IDS) can help decrease smart meters’ energy and resource consumption. Thus, SM-IDS achieves an accuracy of 84.31% with a detection rate of 74.69%. Also, NAN-IDS provides an accuracy of 87.40% and a detection rate of 86.73%.
Current technological advancements in Software Defined Networks (SDN) can provide efficient solutions for smart grids (SGs). An SDN-based SG promises to enhance the efficiency, reliability and sustainability of the communication network. However, new security breaches can be introduced with this adaptation. A layer of defence against insider attacks can be established using machine learning based intrusion detection system (IDS) located on the SDN application layer. Conventional centralised practises, violate the user data privacy aspect, thus distributed or collaborative approaches can be adapted so that attacks can be detected and actions can be taken. This paper proposes a new SDN-based SG architecture, highlighting the existence of IDSs in the SDN application layer. We implemented a new smart meter (SM) collaborative intrusion detection system (SM-IDS), by adapting the split learning methodology. Finally, a comparison of a federated learning and split learning neighbourhood area network (NAN) IDS was made. Numerical results showed, a five class classification accuracy of over 80.3% and F1-score 78.9 for a SM-IDS adapting the split learning technique. Also, the split learning NAN-IDS exhibit an accuracy of over 81.1% and F1-score 79.9.