Abstract
Power grids are essential systems that should resist threats that can be generated by natural or artificial causes. To achieve a resilient system, the potential vulnerabilities have to be studied with detail to provide an impact estimation for the attacks. Also, to predict the elements of the grid that generate the most damage. In previous work, the authors have studied the effects of targeted attacks based in fault strategies under static conditions, however the integration of distributed energy resources makes it necessary evaluating the vulnerability with random variables. This paper presents an analysis of the influence of distributed generation and load variations in the vulnerability of the power system. The vulnerability framework developed in previous research was employed. A time series simulation was carried out with hourly data from irradiance and power demand. Degree, eigenvector, and Katz power traffic centralities were employed to obtain the unsatisfied load derived from the removal of the most central element. The unsatisfied load obtained presented a standard deviation between 0.1682 and 0.1867, and no significant difference was observed between the centralities employed. In addition, a factorial experiment was carried out for two factors: the generation level and the demand level. Twenty five scenarios were simulated for each IEEE test system using five levels of each variable. The VPM maximum variation was 18.8% for levels of generation and load between 50 and 150%. The results indicate that variations in load an generation can significantly affect the vulnerability predictions of the power system under the conditions of this research. The increment in distributed generation can supply the power required to face the outage of important elements in the power system, and the worst case for vulnerability evaluation happens when the generation is minimum and the load is maximum. Therefore, vulnerability studies must account for different operating conditions, and represent UL and VPM as random variables.