Hybrid Control Network Intrusion Detection Systems for Automated Power Distribution Systems

Abstract

Abstract-In this paper, we describe our novel use of network intrusion detection systems (NIDS) for protecting automated distribution systems (ADS) against certain types of cyber attacks in a new way. The novelty consists of using the hybrid control environment rules and model as the baseline for what is normal and what is an anomaly, tailoring the security policies to the physical operation of the system. NIDS sensors in our architecture continuously analyze traffic in the communication medium that comes from embedded controllers, checking if the data and commands exchanged conform to the expected structure of the controllers interactions, and evolution of the system's physical state. Considering its importance in future ADSs, we chose the fault location, isolation and service restoration (FLISR) process as our distribution automation case study for the NIDS deployment. To test our scheme, we emulated the FLISR process using real programmable logic controllers (PLCs) that interact with a simulated physical infrastructure. We used this testbed to examine the capability of our NIDS approach in several attack scenarios. The experimental analysis reveals that our approach is capable of detecting various attacks scenarios including the attacks initiated within the trusted perimeter of the automation network by attackers that have complete knowledge about the communication information exchanged.

A hybrid network IDS for protective digital relays in the power transmission grid

“…In our earlier work [14], we presented a hybrid control NIDS (HC-NIDS) for automated power distribution systems, where we focused on the "fault location, isolation, and service restoration (FLISR)" process. In this paper, our case study is a typical protection mechanism implemented through digital relays used in the power transmission grid.…”

A Hybrid Model for Detecting DDoS Attacks in Wide Area Networks

“…In [13] both AI and Genetic Algorithm (GA) are used in order to have protection from different attacks. Intrusion detection methods in IoT [14], [17] and cyber security mechanisms [15], botnet control traffic detection [16], attack detection in Wireless Sensor Networks (WSNs) [18], Hybrid approach in attack detection in MANET [19] and detection methods for virtual jamming methods [20] are other solutions found in literature. It is understood from the review that there is need for a hybrid approach for detecting different kinds of DDoS attacks in WAN.…”

Monitoring Security of Networked Control Systems: It's the Physics

“…By monitoring the associated DNP3 traffic, we showed that this model allows useful real-time comparisons between relay state and line fault indicators and a simplified topological model of an actual feeder circuit. 13 By developing Bro scripts that are cognizant of the required operation of the FLISR system, we were able to evaluate a variety of attacks against the FLISR system that would cause it to fail to respond appropriately.…”