Using Social Sensors for Detecting Power Outages in the Electrical Utility Industry

Power outages (PO) constitute a serious problem around the world that disrupts our lives in the most unexpected ways. There were 3, 634 power outages reported in 2014, affecting 14.2 million people [3], and over the period of 2008-2014 the US has averaged 2, 987 outages affecting 21.6 million people per year, leading to estimated losses in excess of $150 billion annually [3]. To address this problem, there have been extensive resources dedicated to detecting and reporting power outages with new technologies, such as smart sensors, meters and distribution devices [11]. Utilities’ Outage Management Systems (OMS) vary in their composition of outage detection technologies. These systems typically include both traditional and, so-called, “smart” grid elements as means of power outage detection in the utilities’ coverage regions. Unfortunately, these smart grid technologies are extremely costly, with the total cost being estimated at $338 to $476 billion, and the system will be fully implemented by only 2030 [3]. To address this important problem in the near term and with respect to budgetary constraints, it is necessary to develop alternative approaches to power outage detection. The ubiquity of smart phones and social networks has given rise to an entirely new class of sensor: the human “social sensor” [6]: any individual with a networked device and a social media account can become a “social sensor node” capable of a wide range of functions, including producing data about spontaneous events in real-time functioning either independently or as a collective network. The Electrical Utility industry is uniquely positioned to benefit from incorporating this new class of sensors into their existing Outage Management Systems via existing social networks such as Facebook or Twitter. The data produced by social sensor nodes can be analyzed, modeled, and used to construct a virtual outage detection network for power outage events that could function either independently of or in parallel with utilities’