S. Chanda, R. Dubey, D. Patel, S. Vijay
SAM Analytic Solutions LLC.,
United States
Keywords: animal detection and deterrence solution in electrical substations to avoid power outages
Summary:
Cyber Secure Smart Activated Mesh Solution Based on AI for Distribution Substations Authors: Suraj Chanda, Rajeev Dubey, Dhairya Patel and Samudra Vijay Abstract – The distribution grid’s vast size and complexity adds several layers of challenges to protecting its reliability against system outages caused by several factors. Animal-caused outages in substations are a surprisingly a very common issue for utilities worldwide. Animals, such as squirrels, birds, snakes, raccoons, and even larger creatures, can accidentally interfere with electrical equipment and disrupt service. Impact of these power outages can affect thousands of customers, potential damage to expensive substation equipment and most importantly might pose significant safety risk for both utility workers and the animals. The geographical location of the substations pose an interesting challenge to the utilities – finding a cost effective animal detection & deterrence solution without compromising on the cyber security while connecting into their intra-utility network. This article describes about a cyber-secure smart activated mesh solution based on AI that enables power substations to reduce animal caused outages. The traditional utility approach has been installing animal guards, barriers, fence/mesh and meticulous maintenance of vegetation (tree trimming, grass moving, shrub removal etc.). Reasons including the location of these substations are very remote and buzzing noise attracting the animals make the traditional approach very ineffective to do 24/7 monitoring for animal intrusion. The traditional approach also includes many of utility resources (lineman, field tech etc.). Authors identified that this method is costly and ineffective in a longer run. On the other hand, utilities have stringent budgets to maintain their grid and cannot afford their staff guarding the substations all the time. This motivated the authors to enhance the systems cost-effectiveness by leveraging local trained AI model for animal detection. Smart Activated Mesh (SAM) solution is an animal detection and deterrence system that can be attached to an existing fence or installed standalone with fence posts to detect any animal in the set detection area and deter them without harming the animal. Teams of modular mesh panels are controlled/supervised by a detection box that is installed on one of the team panels. The detection box includes an NDAA camera surveilling the set detection area outside the mesh panels, an AI driven local computing module to analyze camera feed in the detection area/zone, identify the intrusion object and make decision on localized deterrence to activate the deterrence in the respective areas/zones. Considering the system’s requirements to be compatible with SCADA, system utilizes the traditional communicating protocol (DNP3) for all the alarms generated by the image analysis. The authors will discuss development phases; testing efforts associated with solution development; successful commissioning of the first SSAM system; lessons learned; and, provide field operation examples.