Identification of Critical Components and System Vulnerabilities in Network Infrastructures

Electrical grids and ICT systems are network infrastructures that can be modelled as graphs. The integration of electrical, ICT and other energy carrier networks form a system of systems, that can be modelled as a set of interdependent graphs. Graph-theoretical metrics are available to identify critical vertices and edges of a graph and assess its topology. These metrics can be used to identify critical components in network infrastructures. However, an approach based only on a network’s topology neglects any physical properties and limitations and thus, provides limited insight. To improve the informative value of metrics, domain-specific properties should be included. The system of systems is formed by networks from different domains with respective, specific properties and limitations, which poses an additional challenge for the development of metrics applicable to such a system of systems. In this thesis, an approach to analyse interdependent multi-domain network infrastructures with a graph-based approach is to be developed, focusing on electrical and ICT networks as an example. Additional energy carrier networks can be included to prove the transferability and extendibility of the approach. Furthermore, metrics for the identification of critical components and the assessment of vulnerability and resilience of the complete system are to be developed. Graph-theoretical metrics should be used as a starting point and extended to include domain specific properties as well as interdependencies between different domains.

• Critical literature review on state-of-the-art for modelling and analysis of interdependent systems, with a focus on network infrastructures and power systems
• Investigation of existing modelling and analysis approaches and their shortcomings and gaps to include
  • More than two interconnected domains
  • Domain-specific characteristics (e.g., power flows, latency and throughput)
• Development of a modelling approach to overcome these shortcomings
• Development of metrics (multi-domain and/or hybrid: topological and domain-specific) for identification of critical components and system status/vulnerability
• Definition of a relevant example case
• Modelling of the example system and analysis based on the developed metrics

• Knowledge of power systems and automation systems is helpful
• Mathematical background, specifically graph theory is helpful

[1] A. Akbarzadeh und S. Katsikas, „Identifying and Analyzing Dependencies in and among Complex Cyber Physical Systems“, Sensors, Bd. 21, Nr. 5, S. 1685, März 2021, doi: 10.3390/s21051685.

[2] B. Liu, Z. Li, X. Chen, Y. Huang, und X. Liu, „Recognition and Vulnerability Analysis of Key Nodes in Power Grid Based on Complex Network Centrality“, IEEE Trans. Circuits Syst. II, Bd. 65, Nr. 3, S. 346–350, März 2018, doi: 10.1109/TCSII.2017.2705482.

[3] J. V. Milanovic und W. Zhu, „Modeling of Interconnected Critical Infrastructure Systems Using Complex Network Theory“, IEEE Trans. Smart Grid, Bd. 9, Nr. 5, S. 4637–4648, Sep. 2018, doi: 10.1109/TSG.2017.2665646.l