Aging and Rejuvenation Models of Load Changing Attacks in Micro-Grids

Abstract

Recent cyber-attacks in critical infrastructures have highlighted the importance of investigating how to improve Smart-Grids (SG) resiliency. In the future, it is envisioned that grid connected micro-grids would have the ability of operating in 'islanded mode’ in the event of a grid-level failure. In this work, we propose a method for unfolding aging and rejuvenation models into their sequential counterparts to enable the computation of transient state probabilities in the proposed models. We have applied our methodology to one specific security attack scenario and four large campus micro-grids case studies. We have shown how to convert the software aging and rejuvenation, with cycles, to its unfolded counterpart. We then used the unfolded counterpart to support the survivability computation. We were able to analytically evaluate the transient failure probability and the associated Instantaneous Expected Energy Not Supplied metric, for each of the four case studies, from one specific attack. We envision several practical applications of the proposed methodology. First, because the micro-grid model is solved analytically, the approach can be used to support micro-grid engineering optimizations accounting for security intrusions. Second, micro-grid engineers could use the approach to detect security attacks by monitoring for unexpected deviations of the Energy Not Supplied metric.