Robustness Assessment of Interdependent Gas and Electric Power Transmission Systems Against Cyber Threats

By Jose M. Yusta and Jesus Beyza

In 2018 hackers gained access to the control centers of US electric utilities, allowing them to control the status of switches and affect the power flow. IEEE PES Task Force on Understanding, Prediction, Mitigation and Restoration of Cascading Failures has addressed graph theory as an emerging technology on the vulnerability assessment of cascading failures. In this exact sense, we develop a robustness assessment of highly interdependent gas and electric power transmission systems applying a cascading failure methodology based on graph theory in complex networks. A real 400-kV electricity transmission network and an 80-bar natural gas transmission network have been jointly evaluated by means of a graph composed of 2031 nodes and 2154 links. The deliberate disintegration of the tightly coupled networks would only require eliminating approximately 1% of the nodes, confirming the weakness of the energy infrastructures against new cyber threats.

Gas and electricity transmission networks are considered as critical infrastructures, since their operation is essential and does not allow alternative solutions, hence, their disruption or damage would have a serious impact on basic services.

The threats to these infrastructures can be of various types, among others cyber-attacks. The attacks through computer networks against the operators of these facilities grew seven times between 2014 and 2016, and continue so exponentially. Moreover, the energy sector is the industry which suffers one out of every three cyber-attacks in the world. In July 2018, US officials informed the Wall Street Journal that Russian hackers had accessed the control centers of various North American power companies, gaining the capability to change the status of switches of the high-voltage network and interrupt the flow across certain power lines, at the risk of causing important disturbances in the electricity supply. Earlier, in December 2015, Ukraine suffered severe power outages following the intrusion of BlackEnergy malware, while in January 2016 the Israeli electricity network was affected by a massive cyberattack that left its systems inoperative.

On the other hand, there has been increasing concern over the past few years about the growing dependence of electricity infrastructures on natural gas. For example, in just ten years, 67 natural gas generators (total power of 26 GW) have been built in Spain: 25% of the Spanish electricity generation mix currently depends on the supply of natural gas. Unlike the Spanish electricity network, which has progressively expanded to adopt a relatively meshed and highly reliable topological structure, the natural gas transmission network has been developed more recently and does not yet have such an interconnected topology. In addition, the massive penetration of renewable energy will require the construction of new natural gas combined cycle plants that provide sufficient support to ensure security of supply. It is estimated that up to 10 GW of additional capacity will be required in Spain by 2030. With this additional 10 GW, natural gas will constitute 58% of the total firm capacity of the Spanish power system.

At present, it is not possible to consider that gas and electricity transmission networks are isolated, since an event in one infrastructure can have consequences on the other. Combined cycle power plants require a reliable supply of natural gas and, at the same time, many of the natural gas compression stations demand electric power for operation. The problem of interdependence between the two infrastructures must therefore be addressed.

In addition to the efforts made by energy infrastructure operators to implement asset protection planning, researchers have also contributed with the proposal of methodologies for the vulnerability assessment of power systems. These include classic contingency analysis techniques, as power flow, to simulate cascading failures, but also the application of new emerging techniques for structural vulnerability analysis based on complex networks theory. Such techniques have been suggested by the IEEE PES Task Force on Understanding, Prediction, Mitigation and Restoration of Cascading Failures in 2009.

Among the indexes proposed in the scientific literature, metrics based on geodesic efficiency have shown greater correlation with the actual performance of the power transmission networks for the analysis of N-k contingencies, and specifically the geodesic vulnerability index is one of the most innovative. This index measures the functioning of a network against contingencies, normalizing the geodesic efficiency with respect to its base case. This is in turn calculated from the geodesic distance between the node pairs of the graph, after each iteration when a node is removed. Previous research work validated that this index is directly related to the load disconnected in the network under a cascading failure process.

Our research group has developed a cascading failure analysis methodology for interdependent natural gas and electricity coupled systems using the complex network theory. We applied it to the real transmission networks of electricity and natural gas of Spain as a case study. The graph resulting from the 400 kV electrical infrastructure of Spain using only the topology data is composed of a total of 611 nodes and 672 links. In this representation all the main assets have been considered, i.e. power lines, substations, transformers, loads, generators, etc. At the same time, the graph resulting from the topology of the high-pressure natural gas network consists of 1380 nodes and 1402 links, where LNG regasification plants, compression stations, underground storage, gas fields, cross-border connections and transmission-distribution connections have been represented. Finally, the graph resulting from the joint electricity and natural gas grid in Spain is composed of 2031 nodes and 2154 links. Natural gas combined cycle power plants (26) and electric compressors for natural gas (14) act as couplings between the two networks.

Our results show that, faced with malicious attacks, the natural gas system is less robust than the electrical system, due to its less meshed topology. Analyzing the gas and electricity networks in a coupled manner, our research has concluded that the attack on 1% of the nodes would be sufficient to cause complete collapse of the coupled network, using as a cyberattack strategy the combination of the nodal degrees of the independent networks, considering first the sequence of nodes with the greatest connectivity in the electricity network and then those of the natural gas network. These conclusions show the weakness of our modern energy infrastructures under possible malicious cyber-attacks.

Edited by Pardis Khayyer