George Orellana

Yun Wei, Chuanyi Ji, Floyd Galvan et al.

A key objective of the smart grid is to improve reliability of utility services to end users. This requires strengthening resilience of distribution networks that lie at the edge of the grid. However, distribution networks are exposed to external disturbances such as hurricanes and snow storms where electricity service to customers is disrupted repeatedly. External disturbances cause large-scale power failures that are neither well-understood, nor formulated rigorously, nor studied systematically. This work studies resilience of power distribution networks to large-scale disturbances in three aspects. First, a non-stationary random process is derived to characterize an entire life cycle of large-scale failure and recovery. Second, resilience is defined based on the non-stationary random process. Close form analytical expressions are derived under specific large-scale failure scenarios. Third, the non-stationary model and the resilience metric are applied to a real life example of large-scale disruptions due to Hurricane Ike. Real data on large-scale failures from an operational network is used to learn time-varying model parameters and resilience metrics.

Yun Wei, Chuanyi Ji, Floyd Galvan et al.

Smart energy grid is an emerging area for new applications of machine learning in a non-stationary environment. Such a non-stationary environment emerges when large-scale failures occur at power distribution networks due to external disturbances such as hurricanes and severe storms. Power distribution networks lie at the edge of the grid, and are especially vulnerable to external disruptions. Quantifiable approaches are lacking and needed to learn non-stationary behaviors of large-scale failure and recovery of power distribution. This work studies such non-stationary behaviors in three aspects. First, a novel formulation is derived for an entire life cycle of large-scale failure and recovery of power distribution. Second, spatial-temporal models of failure and recovery of power distribution are developed as geo-location based multivariate non-stationary GI(t)/G(t)/Infinity queues. Third, the non-stationary spatial-temporal models identify a small number of parameters to be learned. Learning is applied to two real-life examples of large-scale disruptions. One is from Hurricane Ike, where data from an operational network is exact on failures and recoveries. The other is from Hurricane Sandy, where aggregated data is used for inferring failure and recovery processes at one of the impacted areas. Model parameters are learned using real data. Two findings emerge as results of learning: (a) Failure rates behave similarly at the two different provider networks for two different hurricanes but differently at the geographical regions. (b) Both rapid- and slow-recovery are present for Hurricane Ike but only slow recovery is shown for a regional distribution network from Hurricane Sandy.