The resilience of maritime infrastructures is of increasingly critical importance for society, as ensuring the ongoing operation and performance of these systems is, among other things, necessary for core aspects of mobility, trade and energy generation. The expansion of offshore wind energy over recent years has only heightened the importance of this form of energy supply in the energy mix of the onshore power grid, due to the amount of power it has generated. Awareness of the actual resilience of offshore wind farms is vital in order to stay abreast of these developments. In the past, considerations of this issue tended to focus on individual components or specific safety aims, so established methods and approaches for obtaining a comprehensive overview of resilience are only suitable to a limited extent. The protection of offshore wind farms and possible impairment of the energy supply in the case of attacks or incidents has also hardly been scientifically researched until now.
In its project proposal ‘KPI-based evaluation of the safety and security level in offshore wind farms’, the Institute for the Protection of Maritime Infrastructures aims to develop concepts for optimised situational awareness and the assessment of safety and security status using methods of resilience engineering. This requires a comprehensive understanding of the way in which the system behaves during regular operation, as well as in the event of disruptions. This also includes the emergence of incidents and the impact of any kind of attack.
Generic functional model of a complex system
For this purpose, a generic model of the technical architecture and functional properties of an offshore wind park (OWP) is being created, which is suitable for simulating the system’s behaviour, all with a view to meeting safety and security targets. The associated mathematical modelling of the risk scenarios and the safety and security status requires the comprehensive development of methods to quantify safety- and security-related system states.
In order to be able to make statements about the system status, a network of key performance indicators (KPIs) will be created that is suitable for describing different system properties. Here, a distinction is made between the ‘target’ properties that an OWP should have based on its design and expected risks, and its ‘actual’ properties, which are determined by means of data analysis. Statements about the current degree of resilience achieved by an OWP shall be made on the basis of a comparison between the target and current state. However, the aim is also to detect deviations from nominal behaviour as early as possible, and use these to predict any resulting risks. This creates the sufficient time margin for initiating emergency response measures and limiting damage.
It is thought that the data currently being collected for these purposes is insufficient, which makes it impossible to accurately predict actual vulnerability and resilience. Two essential approaches are being pursued with a view to improving the information value. First, the researchers will investigate how modern data analysis methods can be used to obtain more information using the available sensor architecture, and whether any existing sensors can be used to detect events for which they were not originally designed. As a next step, concepts will be developed for optimised and possibly extended sensor application, which should complete the situation picture.
Finally, data from a range of sources will be evaluated in order to develop a concept for a situation report of the safety and security level of an OWP. This situation picture is intended to provide support to those responsible for OWP safety and security, to initiate proactive and reactive measures as required and as quickly as possible, and thus to ensure the resilience of an OWP. It is expected that KPI-based monitoring offers added value over a classic situation report. One advantage of this approach is that it predicts the behaviour of an OWP in the case of an incident or attack using parameterised system models. As such, it allows for the effects of particular measures to be estimated, while also assessing their priority and level of criticality.
Project duration: October 2018 – September 2020