A Monolithic Cybersecurity Architecture for Power Electronic Systems

TL;DR

This work addresses the vulnerability of power electronic systems to data availability and data integrity attacks by proposing a monolithic cybersecurity architecture (MCA) that embeds semantic principles into the sampling and reconstruction of critical signals at DERs. By exploiting inner control loop dynamics and semantic attributes ($VoI$, $F(t)$, $R(t)$), MCA downscales communications and reconstructs essential information to sustain secondary control objectives under latency, dropouts, TSAs, and FDIAs. The approach is distributed, model-agnostic, training-free, and capable of adapting to dynamic cyber graphs, with validations on a modified IEEE 69-bus and a real-world SCE 47-bus network using OPAL-RT. Results show faster convergence to desired frequency and power sharing, improved robustness to cyber disruptions, and strong scalability, indicating practical potential for resilient PES operation. Future directions include integrating semantic-based demand response and exploring quantum-enabled semantic security to further reduce response times during disturbances.

Abstract

Power electronic systems (PES) face significant threats from various data availability and integrity attacks, significantly affecting the performance of communication networks and power system operation. As a result, several attack detection and reconstruction techniques are deployed, which makes it a costly \& complex cybersecurity operational platform with significant room for incremental extensions for mitigation against future threats. Unlike the said traditional arrangements, our paper introduces a foundational approach by establishing a monolithic cybersecurity architecture (MCA) via incorporating semantic principles into the sampling process for distributed energy resources (DERs). This unified approach concurrently compensates for the intrusion challenges posed by cyber attacks by reconstructing signals using the dynamics of the inner control layer. This reconstruction considers essential semantic attributes, like Priority, Freshness, and Relevance to ensure resilient dynamic performance. Hence, the proposed scheme promises a generalized route to concurrently tackle a global set of cyber attacks in elevating the resilience of PES. Finally, rigorous validation on a modified IEEE 69-bus distribution system and a real-world South California Edison (SCE) 47-bus network, using OPAL-RT under diverse operating conditions, underscores its robustness, model-free design capability, scalability, and adaptability to dynamic cyber graphs and system reconfiguration.

Figures (18)

• Figure 1: (a) Paradigm shift from a complex architecture of cyber-physical resiliency methods to a monolithic cybersecurity architecture (MCA) -- the proposed scheme with semantic sampling offers unified resiliency against data availability and data integrity attacks, (b) Semantic requirements governing the proposed MCA scheme.
• Figure 2: Schematic diagram of the deployment of the proposed scheme (enclosed in dotted blue box) in the existing PES. The semantic information exchange from local error measurements drives the estimation and reconstruction process (to SC) during cyber attacks.
• Figure 3: Modified 69-bus distribution system with nine DERs in partially-connected topology ref16.
• Figure 4: (a) South California Edison's (SCE's) Distributed Energy Resource Interconnection Map (DERiM) ref18; (b) real-world SCE 47-bus network with five DERs ref17 and; (c) cyber topologies T1 and T2.
• Figure 5: The depicted models illustrate cyber attacks on two communicating SCs in terms of (a) latency attack; (b) data dropout; (c) TSA; and (d) FDIA.